Merge to test 20260723 by hadi-aix · Pull Request #1004 · aixplain/aiXplain

hadi-aix · 2026-07-23T17:32:15Z

Merging development into test for the staging promotion on 2026-07-23. development is ahead of test by 33 commits across 9 files.

The three test_issue.py URL assertions hardcoded the production host while Aixplain() resolves backend_url from the BACKEND_URL env var, so they failed whenever .env pointed elsewhere (e.g. test backend). Construct Aixplain with an explicit backend_url to make them deterministic.

* merge to test (#948) * Test to dev sync (#940) * update version * Test (#875) * Test (#807) * Hotfix: Inspectors, v2 fixes, debugger (#801) * ENG-2710 tool in dict form handled (#796) * Eng 2709 SDK 2.0 Various Improvements (#792) * ENG-2709 Added path, developer and host filters * ENG-2709 Removed intermediate steps * ENG-2709 Modified run_async to accept positional query arg * ENG-2709 templated instruction/query * ENG-2709 Fixed instruction templating --------- Co-authored-by: aix-ahmet <ahmet.gunduz@aixplain.com> * ENG-2717 fixes v2 resource repr over path display (#798) * added max iterations and max tokens too agent config (#800) * ENG-2716 Agent status update treatment for notebook envs (#799) * ENG-2716 Agent status update treatment for notebook envs * ENG-2716 experimental flag progress_time_ticks * Add ability to configure rate limits based on input, output or total tokens (#791) * Update api_key.py * Update api * feat(api-key): add token type support for per-asset rate limiting - Add TokenType enum (INPUT, OUTPUT, TOTAL) for configuring which tokens to count for rate limiting purposes - Fix bug where asset limits incorrectly used global tokenType instead of per-asset tokenType - Properly parse tokenType from API response as TokenType enum - Add comprehensive docstrings to TokenType enum and APIKeyUsageLimit class - Move functional API key tests to unit tests with mocked responses - Add unit tests for token type creation, parsing, and serialization Closes ENG-2705 * undo docs --------- Co-authored-by: Hadi <h.nasrallah.c@ncgr.gov.sa> Co-authored-by: ahmetgunduz <ahmet.gunduz44@gmail.com> * safe delete only for dev and test (#804) * ENG-2716 make progress_time_ticks default behaviour (#803) * ENG-2716 Agent status update treatment for notebook envs * ENG-2716 experimental flag progress_time_ticks * ENG-2716 make progress_time_ticks default behaviour * removed inspectors v1, fixed v2 tests (#802) * removed inspectors v1, fixed v2 tests * move import from v1 to v2 --------- Co-authored-by: ahmetgunduz <ahmet.gunduz44@gmail.com> * ENG-2720 Fix team agent save when subagents are saved after team creation (#805) * feat(v2): Add Debugger meta-agent for agent response analysis (#793) * feat(v2): Add Debugger meta-agent for agent response analysis - Add Debugger class in meta_agents.py for analyzing agent runs - Add DebugResult dataclass with analysis property - Add .debug() method to AgentRunResult for chained debugging - Add execution_id extraction from poll URL for enhanced debugging - Export Debugger and DebugResult from v2 module - Register Debugger in Aixplain client context Usage: aix = Aixplain('<api_key>') debugger = aix.Debugger() result = debugger.run('Analyze this agent output...') # Or chained from agent response: response = agent.run('Hello!') debug_result = response.debug() ENG-2635 * Unit test meta agents --------- Co-authored-by: JP Maia <maiajp2305@gmail.com> --------- Co-authored-by: kadirpekel <kadirpekel@gmail.com> Co-authored-by: Zaina Abu Shaban <zaina.abushaban@aixplain.com> Co-authored-by: Hadi Nasrallah <87204330+hadi-aix@users.noreply.github.com> Co-authored-by: Hadi <h.nasrallah.c@ncgr.gov.sa> Co-authored-by: ahmetgunduz <ahmet.gunduz44@gmail.com> Co-authored-by: João Maia <157385649+MaiaJP-AIXplain@users.noreply.github.com> Co-authored-by: JP Maia <maiajp2305@gmail.com> * update functions param * fix evolver tests and add slack integration model id * fix already exists agent error in agent functional test --------- Co-authored-by: kadirpekel <kadirpekel@gmail.com> Co-authored-by: Zaina Abu Shaban <zaina.abushaban@aixplain.com> Co-authored-by: Hadi Nasrallah <87204330+hadi-aix@users.noreply.github.com> Co-authored-by: Hadi <h.nasrallah.c@ncgr.gov.sa> Co-authored-by: ahmetgunduz <ahmet.gunduz44@gmail.com> Co-authored-by: João Maia <157385649+MaiaJP-AIXplain@users.noreply.github.com> Co-authored-by: JP Maia <maiajp2305@gmail.com> * ENG-2722 Fix code/value attribute handling for actions and removed au… (#808) * ENG-2722 Fix code/value attribute handling for actions and removed auth scheme foundation * fix static agent id in tests --------- Co-authored-by: ahmetgunduz <ahmet.gunduz44@gmail.com> * removed limitations other than abort-critical (#811) * removed limitations other than abort-critical * Removed limitations entirely * Make model tool function, supplier, and version optional (#813) * fix: remove INFO severity from inspector - also fixed some unchecked unit tests * ENG-2727-Change-the-inspector-payload-structure (#815) * updated the inspector payload * fix: use snake_case for inspector Python fields, keep camelCase in API payload Renamed maxRetries→max_retries, onExhaust→on_exhaust, assetId→asset_id across InspectorActionConfig, EvaluatorConfig, and EditorConfig. The to_dict()/from_dict() methods still serialize to/from camelCase for the API --------- Co-authored-by: ahmetgunduz <ahmet.gunduz44@gmail.com> * feat(v2): Add Debugger meta-agent for agent response analysis (#814) * feat(v2): Add Debugger meta-agent for agent response analysis - Add Debugger class in meta_agents.py for analyzing agent runs - Add DebugResult dataclass with analysis property - Add .debug() method to AgentRunResult for chained debugging - Add execution_id extraction from poll URL for enhanced debugging - Export Debugger and DebugResult from v2 module - Register Debugger in Aixplain client context Usage: aix = Aixplain('<api_key>') debugger = aix.Debugger() result = debugger.run('Analyze this agent output...') # Or chained from agent response: response = agent.run('Hello!') debug_result = response.debug() ENG-2635 * Unit test meta agents * max tokens --------- Co-authored-by: JP Maia <maiajp2305@gmail.com> * ENG-2726 streaming support for v2 models (#816) * ENG-2726 streaming support for v2 models * ENG-2726 streaming support for v2 models review * ENG-2724 Added v2 apikey resource (#810) * ENG-2724 Added v2 apikey resource * Add tests --------- Co-authored-by: ahmetgunduz <ahmet.gunduz44@gmail.com> * ENG-2725 Adjust run methods so that run uses the v2 endpoint and run_… (#817) * ENG-2725 Adjust run methods so that run uses the v2 endpoint and run_async uses the v1 endpoint * ENG-2725 v1 base url adjusted * ENG-2725 Add tests for connection_type routing and fix completed flag bug --------- Co-authored-by: aix-ahmet <ahmet.gunduz@aixplain.com> Co-authored-by: ahmetgunduz <ahmet.gunduz44@gmail.com> * Get-models-by-name (#674) * model factory get by name * integration factory get by name * add tests --------- Co-authored-by: aix-ahmet <ahmet.gunduz@aixplain.com> * fix model completed * Fix v2 imports from v1 modules that trigger env var validation (#819) When using the v2 SDK with constructor-based API keys (Aixplain(api_key=...)), imports from v1 modules in model.py and utility.py triggered the v1 init chain (aixplain.modules → corpus.py → utils/config.py → validate_api_keys), causing crashes when TEAM_API_KEY env var was not set. - Rewrite _run_async_v1 in v2/model.py to use v2 client directly - Create v2/code_utils.py with self-contained parse_code functions - Add DataType enum to v2/enums.py - Update v2/utility.py to import from v2/code_utils instead of v1 - Add guard test to prevent v1 imports from leaking back into v2 Co-authored-by: ahmetgunduz <ahmet.gunduz44@gmail.com> * Fix models api (#822) * Test (#818) * Packpropagating Connection Description Add (#742) * Test (#737) * Backpropagate the hotfix (#725) * Test (#721) * MCP Deploy error (#681) * Test (#663) * Dev to Test (#654) * Merge to prod (#530) * Merge to test (#246) * Update Finetuner search metadata functional tests (#172) * Downgrade dataclasses-json for compatibility (#170) Co-authored-by: Thiago Castro Ferreira <thiagocastroferreira@ip-192-168-0-4.ec2.internal> * Fix model cost parameters (#179) Co-authored-by: Thiago Castro Ferreira <thiagocastroferreira@ip-192-168-0-4.ec2.internal> * Treat label URLs (#176) Co-authored-by: Thiago Castro Ferreira <thiagocastroferreira@ip-192-168-0-4.ec2.internal> * Add new metric test (#181) * Add new metric test * Enable testing new pipeline executor --------- Co-authored-by: Thiago Castro Ferreira <thiago@aixplain.com> * LLMModel class and parameters (#184) * LLMModel class and parameters * Change in the documentation * Changing LLMModel for LLM * Remove frequency penalty --------- Co-authored-by: Thiago Castro Ferreira <thiago@aixplain.com> * Gpus (#185) * Release. (#141) * Merge dev to test (#107) * Create bounds for FineTune hyperparameters (#103) * Test bound to hyperparameters * Update finetune llm hyperparameters * Remove option to use PEFT, always on use now * Fixing pipeline general asset test (#106) --------- Co-authored-by: Thiago Castro Ferreira <85182544+thiago-aixplain@users.noreply.github.com> * Development to Test (#109) * Create bounds for FineTune hyperparameters (#103) * Test bound to hyperparameters * Update finetune llm hyperparameters * Remove option to use PEFT, always on use now * Fixing pipeline general asset test (#106) --------- Co-authored-by: Lucas Pavanelli <86805709+lucas-aixplain@users.noreply.github.com> * Merge to test (#111) * Create bounds for FineTune hyperparameters (#103) * Test bound to hyperparameters * Update finetune llm hyperparameters * Remove option to use PEFT, always on use now * Fixing pipeline general asset test (#106) --------- Co-authored-by: Lucas Pavanelli <86805709+lucas-aixplain@users.noreply.github.com> Co-authored-by: Thiago Castro Ferreira <85182544+thiago-aixplain@users.noreply.github.com> * Merge dev to test (#113) * Create bounds for FineTune hyperparameters (#103) * Test bound to hyperparameters * Update finetune llm hyperparameters * Remove option to use PEFT, always on use now * Fixing pipeline general asset test (#106) * Update Finetuner functional tests (#112) --------- Co-authored-by: Thiago Castro Ferreira <85182544+thiago-aixplain@users.noreply.github.com> * Hf deployment test (#114) * Started adding Hugging Face deployment to aiXplain SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added model status function to SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updating Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updated CLI Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Adding CLI Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Corrected request error Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Clearing out unnecessary information in return Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Simplifying status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests and correcting tokens Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added bad repo ID test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Finished rough draft of tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Fixing hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Correcting first test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Testing Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding config Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added user doc Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added gated model test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> --------- Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Hf deployment test (#118) * Started adding Hugging Face deployment to aiXplain SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added model status function to SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updating Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updated CLI Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Create bounds for FineTune hyperparameters (#103) * Test bound to hyperparameters * Update finetune llm hyperparameters * Remove option to use PEFT, always on use now * Fixing pipeline general asset test (#106) * Adding CLI Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Corrected request error Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Clearing out unnecessary information in return Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Simplifying status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests and correcting tokens Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added bad repo ID test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Finished rough draft of tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Fixing hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Correcting first test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Testing Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding config Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added user doc Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added gated model test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Update Finetuner functional tests (#112) * Hf deployment test (#115) * Started adding Hugging Face deployment to aiXplain SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added model status function to SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updating Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updated CLI Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Adding CLI Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Corrected request error Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Clearing out unnecessary information in return Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Simplifying status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests and correcting tokens Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added bad repo ID test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Finished rough draft of tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Fixing hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Correcting first test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Testing Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding config Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added user doc Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added gated model test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> --------- Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Adding HF token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> --------- Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> Signed-off-by: mikelam-us <michael.lam@aixplain.com> Co-authored-by: Lucas Pavanelli <86805709+lucas-aixplain@users.noreply.github.com> Co-authored-by: Thiago Castro Ferreira <85182544+thiago-aixplain@users.noreply.github.com> * Merge to test (#124) * Create bounds for FineTune hyperparameters (#103) * Test bound to hyperparameters * Update finetune llm hyperparameters * Remove option to use PEFT, always on use now * Fixing pipeline general asset test (#106) * Update Finetuner functional tests (#112) * Hf deployment test (#115) * Started adding Hugging Face deployment to aiXplain SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added model status function to SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updating Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updated CLI Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Adding CLI Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Corrected request error Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Clearing out unnecessary information in return Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Simplifying status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests and correcting tokens Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added bad repo ID test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Finished rough draft of tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Fixing hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Correcting first test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Testing Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding config Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added user doc Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added gated model test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> --------- Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Hf deployment test (#117) * Started adding Hugging Face deployment to aiXplain SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added model status function to SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updating Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updated CLI Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Adding CLI Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Corrected request error Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Clearing out unnecessary information in return Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Simplifying status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests and correcting tokens Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added bad repo ID test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Finished rough draft of tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Fixing hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Correcting first test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Testing Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding config Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added user doc Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added gated model test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Merge dev to test (#113) * Create bounds for FineTune hyperparameters (#103) * Test bound to hyperparameters * Update finetune llm hyperparameters * Remove option to use PEFT, always on use now * Fixing pipeline general asset test (#106) * Update Finetuner functional tests (#112) --------- Co-authored-by: Thiago Castro Ferreira <85182544+thiago-aixplain@users.noreply.github.com> * Hf deployment test (#114) * Started adding Hugging Face deployment to aiXplain SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added model status function to SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updating Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updated CLI Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Adding CLI Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Corrected request error Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Clearing out unnecessary information in return Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Simplifying status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests and correcting tokens Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added bad repo ID test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Finished rough draft of tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Fixing hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Correcting first test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Testing Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding config Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added user doc Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added gated model test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> --------- Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Adding HF token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> --------- Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> Signed-off-by: mikelam-us <michael.lam@aixplain.com> Co-authored-by: Lucas Pavanelli <86805709+lucas-aixplain@users.noreply.github.com> Co-authored-by: Thiago Castro Ferreira <85182544+thiago-aixplain@users.noreply.github.com> * Do not download textual URLs (#120) * Do not download textual URLs * Treat as string --------- Co-authored-by: Thiago Castro Ferreira <thiagocastroferreira@ip-192-168-0-3.ec2.internal> * Enable api key parameter in data asset creation (#122) Co-authored-by: Thiago Castro Ferreira <thiagocastroferreira@ip-192-168-0-3.ec2.internal> --------- Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> Signed-off-by: mikelam-us <michael.lam@aixplain.com> Co-authored-by: Lucas Pavanelli <86805709+lucas-aixplain@users.noreply.github.com> Co-authored-by: Thiago Castro Ferreira <85182544+thiago-aixplain@users.noreply.github.com> Co-authored-by: mikelam-us-aixplain <131073216+mikelam-us-aixplain@users.noreply.github.com> Co-authored-by: Thiago Castro Ferreira <thiagocastroferreira@ip-192-168-0-3.ec2.internal> * Merge dev to test (#126) * Create bounds for FineTune hyperparameters (#103) * Test bound to hyperparameters * Update finetune llm hyperparameters * Remove option to use PEFT, always on use now * Fixing pipeline general asset test (#106) * Update Finetuner functional tests (#112) * Hf deployment test (#115) * Started adding Hugging Face deployment to aiXplain SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added model status function to SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updating Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updated CLI Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Adding CLI Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Corrected request error Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Clearing out unnecessary information in return Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Simplifying status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests and correcting tokens Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added bad repo ID test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Finished rough draft of tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Fixing hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Correcting first test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Testing Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding config Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added user doc Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added gated model test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> --------- Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Hf deployment test (#117) * Started adding Hugging Face deployment to aiXplain SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added model status function to SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updating Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updated CLI Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Adding CLI Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Corrected request error Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Clearing out unnecessary information in return Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Simplifying status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests and correcting tokens Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added bad repo ID test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Finished rough draft of tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Fixing hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Correcting first test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Testing Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding config Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added user doc Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added gated model test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Merge dev to test (#113) * Create bounds for FineTune hyperparameters (#103) * Test bound to hyperparameters * Update finetune llm hyperparameters * Remove option to use PEFT, always on use now * Fixing pipeline general asset test (#106) * Update Finetuner functional tests (#112) --------- Co-authored-by: Thiago Castro Ferreira <85182544+thiago-aixplain@users.noreply.github.com> * Hf deployment test (#114) * Started adding Hugging Face deployment to aiXplain SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added model status function to SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updating Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updated CLI Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Adding CLI Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Corrected request error Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Clearing out unnecessary information in return Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Simplifying status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests and correcting tokens Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added bad repo ID test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Finished rough draft of tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Fixing hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Correcting first test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Testing Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding config Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added user doc Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added gated model test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> --------- Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Adding HF token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> --------- Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> Signed-off-by: mikelam-us <michael.lam@aixplain.com> Co-authored-by: Lucas Pavanelli <86805709+lucas-aixplain@users.noreply.github.com> Co-authored-by: Thiago Castro Ferreira <85182544+thiago-aixplain@users.noreply.github.com> * Do not download textual URLs (#120) * Do not download textual URLs * Treat as string --------- Co-authored-by: Thiago Castro Ferreira <thiagocastroferreira@ip-192-168-0-3.ec2.internal> * Enable api key parameter in data asset creation (#122) Co-authored-by: Thiago Castro Ferreira <thiagocastroferreira@ip-192-168-0-3.ec2.internal> * Update Finetuner hyperparameters (#125) * Update Finetuner hyperparameters * Change hyperparameters error message --------- Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> Signed-off-by: mikelam-us <michael.lam@aixplain.com> Co-authored-by: Thiago Castro Ferreira <85182544+thiago-aixplain@users.noreply.github.com> Co-authored-by: mikelam-us-aixplain <131073216+mikelam-us-aixplain@users.noreply.github.com> Co-authored-by: Thiago Castro Ferreira <thiagocastroferreira@ip-192-168-0-3.ec2.internal> * Merge dev to test (#129) * Create bounds for FineTune hyperparameters (#103) * Test bound to hyperparameters * Update finetune llm hyperparameters * Remove option to use PEFT, always on use now * Fixing pipeline general asset test (#106) * Update Finetuner functional tests (#112) * Hf deployment test (#115) * Started adding Hugging Face deployment to aiXplain SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added model status function to SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updating Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updated CLI Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Adding CLI Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Corrected request error Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Clearing out unnecessary information in return Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Simplifying status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests and correcting tokens Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added bad repo ID test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Finished rough draft of tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Fixing hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Correcting first test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Testing Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding config Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added user doc Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added gated model test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> --------- Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Hf deployment test (#117) * Started adding Hugging Face deployment to aiXplain SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added model status function to SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updating Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updated CLI Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Adding CLI Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Corrected request error Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Clearing out unnecessary information in return Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Simplifying status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests and correcting tokens Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added bad repo ID test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Finished rough draft of tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Fixing hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Correcting first test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Testing Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding config Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added user doc Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added gated model test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Merge dev to test (#113) * Create bounds for FineTune hyperparameters (#103) * Test bound to hyperparameters * Update finetune llm hyperparameters * Remove option to use PEFT, always on use now * Fixing pipeline general asset test (#106) * Update Finetuner functional tests (#112) --------- Co-authored-by: Thiago Castro Ferreira <85182544+thiago-aixplain@users.noreply.github.com> * Hf deployment test (#114) * Started adding Hugging Face deployment to aiXplain SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added model status function to SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updating Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updated CLI Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Adding CLI Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Corrected request error Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Clearing out unnecessary information in return Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Simplifying status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests and correcting tokens Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added bad repo ID test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Finished rough draft of tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Fixing hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Correcting first test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Testing Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding config Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added user doc Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added gated model test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> --------- Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Adding HF token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> --------- Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> Signed-off-by: mikelam-us <michael.lam@aixplain.com> Co-authored-by: Lucas Pavanelli <86805709+lucas-aixplain@users.noreply.github.com> Co-authored-by: Thiago Castro Ferreira <85182544+thiago-aixplain@users.noreply.github.com> * Do not download textual URLs (#120) * Do not download textual URLs * Treat as string --------- Co-authored-by: Thiago Castro Ferreira <thiagocastroferreira@ip-192-168-0-3.ec2.internal> * Enable api key parameter in data asset creation (#122) Co-authored-by: Thiago Castro Ferreira <thiagocastroferreira@ip-192-168-0-3.ec2.internal> * Update Finetuner hyperparameters (#125) * Update Finetuner hyperparameters * Change hyperparameters error message * Add new LLMs finetuner models (mistral and solar) (#128) --------- Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> Signed-off-by: mikelam-us <michael.lam@aixplain.com> Co-authored-by: Thiago Castro Ferreira <85182544+thiago-aixplain@users.noreply.github.com> Co-authored-by: mikelam-us-aixplain <131073216+mikelam-us-aixplain@users.noreply.github.com> Co-authored-by: Thiago Castro Ferreira <thiagocastroferreira@ip-192-168-0-3.ec2.internal> * Merge to test (#135) * Create bounds for FineTune hyperparameters (#103) * Test bound to hyperparameters * Update finetune llm hyperparameters * Remove option to use PEFT, always on use now * Fixing pipeline general asset test (#106) * Update Finetuner functional tests (#112) * Hf deployment test (#115) * Started adding Hugging Face deployment to aiXplain SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added model status function to SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updating Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updated CLI Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Adding CLI Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Corrected request error Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Clearing out unnecessary information in return Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Simplifying status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests and correcting tokens Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added bad repo ID test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Finished rough draft of tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Fixing hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Correcting first test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Testing Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding config Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added user doc Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added gated model test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> --------- Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Hf deployment test (#117) * Started adding Hugging Face deployment to aiXplain SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added model status function to SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updating Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updated CLI Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Adding CLI Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Corrected request error Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Clearing out unnecessary information in return Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Simplifying status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests and correcting tokens Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added bad repo ID test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Finished rough draft of tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Fixing hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Correcting first test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Testing Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding config Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added user doc Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added gated model test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Merge dev to test (#113) * Create bounds for FineTune hyperparameters (#103) * Test bound to hyperparameters * Update finetune llm hyperparameters * Remove option to use PEFT, always on use now * Fixing pipeline general asset test (#106) * Update Finetuner functional tests (#112) --------- Co-authored-by: Thiago Castro Ferreira <85182544+thiago-aixplain@users.noreply.github.com> * Hf deployment test (#114) * Started adding Hugging Face deployment to aiXplain SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added model status function to SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updating Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updated CLI Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Adding CLI Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Corrected request error Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Clearing out unnecessary information in return Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Simplifying status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests and correcting tokens Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added bad repo ID test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Finished rough draft of tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Fixing hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Correcting first test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Testing Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding config Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added user doc Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added gated model test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> --------- Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Adding HF token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> --------- Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> Signed-off-by: mikelam-us <michael.lam@aixplain.com> Co-authored-by: Lucas Pavanelli <86805709+lucas-aixplain@users.noreply.github.com> Co-authored-by: Thiago Castro Ferreira <85182544+thiago-aixplain@users.noreply.github.com> * Do not download textual URLs (#120) * Do not download textual URLs * Treat as string --------- Co-authored-by: Thiago Castro Ferreira <thiagocastroferreira@ip-192-168-0-3.ec2.internal> * Enable api key parameter in data asset creation (#122) Co-authored-by: Thiago Castro Ferreira <thiagocastroferreira@ip-192-168-0-3.ec2.internal> * Update Finetuner hyperparameters (#125) * Update Finetuner hyperparameters * Change hyperparameters error message * Add new LLMs finetuner models (mistral and solar) (#128) * Enabling dataset ID and model ID as parameters for finetuner creation (#131) Co-authored-by: Thiago Castro Ferreira <thiagocastroferreira@ip-192-168-0-4.ec2.internal> * Fix supplier representation of a model (#132) * Fix supplier representation of a model * Fixing parameter typing --------- Co-authored-by: Thiago Castro Ferreira <thiagocastroferreira@ip-192-168-0-4.ec2.internal> * Fixing indentation in documentation sample code (#134) Co-authored-by: Thiago Castro Ferreira <thiagocastroferreira@ip-192-168-0-4.ec2.internal> --------- Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> Signed-off-by: mikelam-us <michael.lam@aixplain.com> Co-authored-by: Lucas Pavanelli <86805709+lucas-aixplain@users.noreply.github.com> Co-authored-by: Thiago Castro Ferreira <85182544+thiago-aixplain@users.noreply.github.com> Co-authored-by: mikelam-us-aixplain <131073216+mikelam-us-aixplain@users.noreply.github.com> Co-authored-by: Thiago Castro Ferreira <thiagocastroferreira@ip-192-168-0-3.ec2.internal> Co-authored-by: Thiago Castro Ferreira <thiagocastroferreira@ip-192-168-0-4.ec2.internal> * Merge dev to test (#137) * Create bounds for FineTune hyperparameters (#103) * Test bound to hyperparameters * Update finetune llm hyperparameters * Remove option to use PEFT, always on use now * Fixing pipeline general asset test (#106) * Update Finetuner functional tests (#112) * Hf deployment test (#115) * Started adding Hugging Face deployment to aiXplain SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added model status function to SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updating Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updated CLI Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Adding CLI Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Corrected request error Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Clearing out unnecessary information in return Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Simplifying status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests and correcting tokens Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added bad repo ID test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Finished rough draft of tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Fixing hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Correcting first test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Testing Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding config Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added user doc Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added gated model test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> --------- Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Hf deployment test (#117) * Started adding Hugging Face deployment to aiXplain SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added model status function to SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updating Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updated CLI Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Adding CLI Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Corrected request error Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Clearing out unnecessary information in return Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Simplifying status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests and correcting tokens Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added bad repo ID test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Finished rough draft of tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Fixing hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Correcting first test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Testing Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding config Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added user doc Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added gated model test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Merge dev to test (#113) * Create bounds for FineTune hyperparameters (#103) * Test bound to hyperparameters * Update finetune llm hyperparameters * Remove option to use PEFT, always on use now * Fixing pipeline general asset test (#106) * Update Finetuner functional tests (#112) --------- Co-authored-by: Thiago Castro Ferreira <85182544+thiago-aixplain@users.noreply.github.com> * Hf deployment test (#114) * Started adding Hugging Face deployment to aiXplain SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added model status function to SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updating Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updated CLI Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Adding CLI Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Corrected request error Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Clearing out unnecessary information in return Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Simplifying status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests and correcting tokens Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added bad repo ID test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Finished rough draft of tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Fixing hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Correcting first test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Testing Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding config Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added user doc Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added gated model test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> --------- Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Adding HF token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> --------- Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> Signed-off-by: mikelam-us <michael.lam@aixplain.com> Co-authored-by: Lucas Pavanelli <86805709+lucas-aixplain@users.noreply.github.com> Co-authored-by: Thiago Castro Ferreira <85182544+thiago-aixplain@users.noreply.github.com> * Do not download textual URLs (#120) * Do not download textual URLs * Treat as string --------- Co-authored-by: Thiago Castro Ferreira <thiagocastroferreira@ip-192-168-0-3.ec2.internal> * Enable api key parameter in data asset creation (#122) Co-authored-by: Thiago Castro Ferreira <thiagocastroferreira@ip-192-168-0-3.ec2.internal> * Update Finetuner hyperparameters (#125) * Update Finetuner hyperparameters * Change hyperparameters error message * Add new LLMs finetuner models (mistral and solar) (#128) * Enabling dataset ID and model ID as parameters for finetuner creation (#131) Co-authored-by: Thiago Castro Ferreira <thiagocastroferreira@ip-192-168-0-4.ec2.internal> * Fix supplier representation of a model (#132) * Fix supplier representation of a model * Fixing parameter typing --------- Co-authored-by: Thiago Castro Ferreira <thiagocastroferreira@ip-192-168-0-4.ec2.internal> * Fixing indentation in documentation sample code (#134) Co-authored-by: Thiago Castro Ferreira <thiagocastroferreira@ip-192-168-0-4.ec2.internal> * Update FineTune unit and functional tests (#136) --------- Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> Signed-off-by: mikelam-us <michael.lam@aixplain.com> Co-authored-by: Thiago Castro Ferreira <85182544+thiago-aixplain@users.noreply.github.com> Co-authored-by: mikelam-us-aixplain <131073216+mikelam-us-aixplain@users.noreply.github.com> Co-authored-by: Thiago Castro Ferreira <thiagocastroferreira@ip-192-168-0-3.ec2.internal> Co-authored-by: Thiago Castro Ferreira <thiagocastroferreira@ip-192-168-0-4.ec2.internal> --------- Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> Signed-off-by: mikelam-us <michael.lam@aixplain.com> Co-authored-by: Lucas Pavanelli <86805709+lucas-aixplain@users.noreply.github.com> Co-authored-by: Thiago Castro Ferreira <85182544+thiago-aixplain@users.noreply.github.com> Co-authored-by: mikelam-us-aixplain <131073216+mikelam-us-aixplain@users.noreply.github.com> Co-authored-by: Thiago Castro Ferreira <thiagocastroferreira@ip-192-168-0-3.ec2.internal> Co-authored-by: Thiago Castro Ferreira <thiagocastroferreira@ip-192-168-0-4.ec2.internal> * Merge to prod. (#152) * Merge dev to test (#107) * Create bounds for FineTune hyperparameters (#103) * Test bound to hyperparameters * Update finetune llm hyperparameters * Remove option to use PEFT, always on use now * Fixing pipeline general asset test (#106) --------- Co-authored-by: Thiago Castro Ferreira <85182544+thiago-aixplain@users.noreply.github.com> * Development to Test (#109) * Create bounds for FineTune hyperparameters (#103) * Test bound to hyperparameters * Update finetune llm hyperparameters * Remove option to use PEFT, always on use now * Fixing pipeline general asset test (#106) --------- Co-authored-by: Lucas Pavanelli <86805709+lucas-aixplain@users.noreply.github.com> * Merge to test (#111) * Create bounds for FineTune hyperparameters (#103) * Test bound to hyperparameters * Update finetune llm hyperparameters * Remove option to use PEFT, always on use now * Fixing pipeline general asset test (#106) --------- Co-authored-by: Lucas Pavanelli <86805709+lucas-aixplain@users.noreply.github.com> Co-authored-by: Thiago Castro Ferreira <85182544+thiago-aixplain@users.noreply.github.com> * Merge dev to test (#113) * Create bounds for FineTune hyperparameters (#103) * Test bound to hyperparameters * Update finetune llm hyperparameters * Remove option to use PEFT, always on use now * Fixing pipeline general asset test (#106) * Update Finetuner functional tests (#112) --------- Co-authored-by: Thiago Castro Ferreira <85182544+thiago-aixplain@users.noreply.github.com> * Hf deployment test (#114) * Started adding Hugging Face deployment to aiXplain SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added model status function to SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updating Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updated CLI Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Adding CLI Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Corrected request error Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Clearing out unnecessary information in return Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Simplifying status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests and correcting tokens Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added bad repo ID test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Finished rough draft of tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Fixing hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Correcting first test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Testing Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding config Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added user doc Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added gated model test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> --------- Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Hf deployment test (#118) * Started adding Hugging Face deployment to aiXplain SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added model status function to SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updating Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updated CLI Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Create bounds for FineTune hyperparameters (#103) * Test bound to hyperparameters * Update finetune llm hyperparameters * Remove option to use PEFT, always on use now * Fixing pipeline general asset test (#106) * Adding CLI Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Corrected request error Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Clearing out unnecessary information in return Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Simplifying status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests and correcting tokens Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added bad repo ID test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Finished rough draft of tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Fixing hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Correcting first test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Testing Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding config Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added user doc Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added gated model test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Update Finetuner functional tests (#112) * Hf deployment test (#115) * Started adding Hugging Face deployment to aiXplain SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added model status function to SDK Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updating Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Updated CLI Signed-off-by: mikelam-us <michael.lam@aixplain.com> * Adding CLI Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Corrected request error Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Clearing out unnecessary information in return Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Simplifying status Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests and correcting tokens Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Added bad repo ID test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Finished rough draft of tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding tests Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Fixing hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Adding hf token Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Correcting first test Signed-off-by: mikelam-us-aixplain <michael.lam@aixplain.com> * Testing Signed-off-by: mikela…

… payload is provided A vision call via run(data=[...]) was rejected by _validate_params because the model declares text/prompt as required inputs, even though the data payload carries the full conversation. The ValueError surfaced before the request reached the backend, forcing callers into manual build/submit paths that skip polling and return an IN_PROGRESS result with empty output. Skip the missing-required check for any param when a non-None data payload is present; type validation of explicitly provided params is unchanged.

…ds (#863) * Add backend-managed session CRUD, messages, reactions, and file uploads Introduce Session and SessionMessage resources with full lifecycle support (create, get, list, update, delete), message operations (add, list, get, delete), reactions (like/dislike/clear), and file upload attachments. Deprecate client-side generate_session_id() in favor of create_session(). Includes 87 unit tests and functional test suite. * remove generate_session_id * Feat/agent run via session (#918) * feat(agent): add via_session=True opt-in path on Agent.run + ExecutionConfig on Session Adds an opt-in `via_session=True` flag to `Agent.run()` that routes the run through the new sessions+messages endpoints while preserving the legacy `AgentRunResult` shape (data.steps, execution_stats, used_credits, run_time). Flow when via_session=True: 1. Get-or-create a Session carrying the supplied executionConfig (execution_params, criteria, evolve, identifier, run_response_generation). 2. POST a role="user" message via session.add_message — backend triggers the agent run with the session's executionConfig. 3. Pull request_id off the user message and poll the legacy /sdk/agents/{request_id}/result endpoint via self.sync_poll, returning a fully populated AgentRunResult. The legacy default (via_session=False) is unchanged. run_async + via_session raises NotImplementedError for now (sync-only). Legacy-only kwargs (tasks, prompt, inspectors, history, variables, allow_history_and_session_id) raise ValueError under via_session=True; the message-trigger contract on the backend doesn't carry them yet. Progress display (progress_format/verbosity/truncate) works identically on both paths via shared _start/_finish_progress_tracker helpers. Session resource gains an `execution_config` field (camelCase `executionConfig`); ExecutionConfig dataclass normalizes snake/camelCase execution_params and is also reachable via Agent.create_session shortcuts. Tests: 601 v2 unit tests pass (was 590); new file test_agent_via_session.py covers session create vs reuse, missing requestId guard, legacy-kwarg rejection, run_async hard-error, and a regression-style legacy-default smoke. test_session.py adds ExecutionConfig serialization/coercion and Session payload coverage. * Merge origin/development into feat/agent-run-via-session Brings in 60 commits from development, including: - ENG-2880: Deprecate `subagents` param, replace with `agents` - Pre-built inspectors support (#913) - New v2 RLM resource - v2 Actions / Inputs hierarchy - Default LLM bumped to GPT-5.4 (#888) - `user_info_utils` for client-side run metadata Conflict resolutions (additive on both sides): - aixplain/v2/__init__.py: keep both Session enum exports and the new Actions/Inputs hierarchy in __all__ - aixplain/v2/core.py: keep both Session and RLM imports / TypeVars / class attrs / __init__ assignments - .gitignore: keep both .claude/ and .cursor* entries Fixes applied as part of the merge: - aixplain/utils/user_info_utils.py: broaden the _fetch_ipinfo except clause from `(RequestException, ValueError)` to `Exception`. The helper makes a network call to ipinfo.io during agent.run(); under requests_mock the unmocked URL raises NoMockAddress (not a subclass of RequestException), which propagated and broke v1 agent tests. This is auxiliary metadata — any failure should degrade silently. - .pre-commit-config.yaml: add explicit additional_dependencies for the pytest-check hook so its isolated venv has the project's runtime deps (Babel, pydantic, pandas, etc.). Without this, the hook venv was missing Babel after the merge. Tests: 1104/1104 unit tests pass after merge + fixes. * Add TYPE_CHECKING to agent.py * Forward attachments and files to user messages * Restore generate_session_id deprecation shim and honor per-run overrides on reused sessions --------- Co-authored-by: Abdelrahman Elsheikh <abdelrahman.el-sheikh@aixplain.com>

* docs: swap README logo to current aixplain brand mark Replace the deprecated "aiXplain" wordmark with the current brand lockup (iron-butterfly + lowercase "aixplain"). Adds light/dark SVG variants and points the README <picture> block at them. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * docs: refresh SDK v2 README — guardrails register, correct system agents, simpler snippets - Lead with developer value; frame governance as "guardrails/control you don't have to build" instead of enterprise RBAC up top (RBAC/SOC2 kept in Data handling) - Correct micro-agents to Planner/Orchestrator/Inspector/Bodyguard (remove Responder; fix Mentalist→Planner); update execution-order diagram - Drop the outdated team-agent workflow image - Remove uuid4 dependency from quick-start snippets (static agent names) Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * docs: fix architecture SVG naming to canon - "AIXPLAIN AGENTIC OS" → "aixplain Agentic OS" (lowercase brand) - "AGENTENGINE" → "AGENT ENGINE", "ASSETSERVING" → "ASSET SERVING" (spaced names) - "Retrieval Engine" → "aiR" (canonical service name) Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * docs: redesign architecture diagram + embed in README - Rebuild the Agentic OS architecture SVG on-brand: governed execution loop as the focal point (Goal → Planner → Orchestrator → [Inspector · Bodyguard] → Action → Result, replan-on-failure), six kernel services, Marketplace + BYO, Evolver feedback, and a Cloud/On-Prem/Edge/Air-gapped deploy band - Add control axes (your data · cost · independence); drop rainbow arrows and drop shadows; one orange focal; white panel reads in light + dark - Embed the diagram in the README Agentic OS section Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * docs: trim Core concepts — drop agent-internals rows (now in the diagram) Keep the five SDK-object concepts (Agent, Team agent, Tool, Model, Integration); remove the Micro-agents/Meta-agent rows (shown in the architecture diagram and a maintenance/drift surface). Point to the diagram + docs instead. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * docs: remove console.aixplain.com links (stale URL) Drop the PAYG API key badge and delink the quick-start "aixplain account" reference; the console URL was stale. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * docs: set SDK v1 sunset to August 1, 2026 (and SDK 1 → SDK v1) Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * docs: say 'infrastructure' not 'kernel' (external voice rule) kernel is internal metaphor only; external surfaces use infrastructure/runtime. Diagram kicker RUNS ON THE KERNEL -> RUNS ON THE INFRASTRUCTURE; README alt text updated. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * docs: update Agentic OS architecture diagram (polished) Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * docs: add control axes (your data / cost / independence) to README intro Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * docs: tighten README intro (headline + 3 short paragraphs) Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * docs: de-stack 'Agentic OS' + drop 'platform'; deploy modes = cloud/on-prem/edge/local; cost free on local - Agentic OS section: reword prose to not repeat 'Agentic OS' and drop 'platform' (voice) - Deploy modes: cloud, on-prem, edge, local (air-gapped is a property of on-prem, kept under it in Data handling) - Cost axis: free on local models and tools, pay as you go in the cloud Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * docs: simplify README — drop Why-aixplain table + Core concepts, rename to Marketplace (SDK/API/MCP outlets), tighten Data handling - Why aixplain: keep the bullets, cut the vs-frameworks comparison table - Remove Core concepts (redundant with quick start + diagram + docs) - 'MCP Server Marketplace' -> 'Marketplace'; note every asset has SDK, API, and MCP outlets - Data handling: 6 bullets -> 3; air-gapped framed as an option of on-prem or local Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * docs: rename 'Agentic OS' section to 'How it works' (de-stack with diagram title) Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> * docs: sync docs/README to the refreshed root README (canon-clean, correct asset paths) Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> --------- Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>

* added warning for saving agent with integration * added name comment for clarification

* wrapped non-json * removed debug

…ads (#967) * feat(v2): override aiXplain tool parameters in agent create/run payloads Allow per-tool parameter overrides to be supplied in both Agent create and run payloads using a shared user-facing shape: ``[{"id": <tool_id>, "parameters": {...}}]`` (plain string ids also accepted). The SDK normalizes these into the platform's ``[{name, value}]`` shape and forwards them; the backend merges by tool id, overriding the agent's persisted per-tool parameters. - Add ``tools`` to AgentRunParams for run-time overrides. - Extract shared ``_normalize_tool_for_api`` used by both create and run. - Support dict-shaped per-tool ``parameters`` ({key: value}) alongside the existing as_tool() snapshot list shape. Closes #966 * feat(v2): session-level tool parameters with per-message override The via_session=True run path silently dropped the run-time tools override. Add session-level tool params on create_session (executionConfig.tools) and a per-message tools override on Session.add_message, and route the run-time tools kwarg through _run_via_session into add_message so it is no longer dropped. Per-message tools win over session-level params by tool id (backend merge). Single-shot create/run behavior is unchanged. Reuses _normalize_tool_for_api for the shared {id, parameters:[{name,value}]} wire shape. * fix(v2): resolve tool type for string-id and {id, parameters} shapes The create payload requires a 'type' on each tools entry, but _normalize_tool_for_api emitted none for the bare string-id and {id, parameters} shapes — so agent create failed with 'tools.0.type should not be empty'. Resolve the asset by id (Tool, then Model) to its as_tool() snapshot for the correct type, then overlay the per-tool parameters override. Dicts that already carry a type pass through unchanged; unresolvable ids degrade to the prior bare entry. Adds tests.

This reverts commit 0cc1e52.

* ENG-3191:check url before polling * ENG-3191: fix id in test

* final shape doesn't include url * ensure same behavior on v2

…s fetch/save (#982) Pins the 0.2.43/0.2.44 regression where Agent.get() fell back to DEFAULT_LLM and any save() overwrote the agent's configured LLM.

A BaseModel-class expected_output was popped from the save payload as "runtime-only", so the backend stored null; the create-response re-hydration then wiped the local value and subsequent runs sent expectedOutput: null, losing the JSON contract. Serialize the model's JSON schema into the save payload instead, with regression tests covering save payload shape, post-save re-hydration, and the run payload after save.

…954) * feat(v2): send per-run budget to agents via execution_params (#953) Add a run-time Budget for agent runs that serializes into executionParams.budget on the wire. - New Budget dataclass (max_cost / max_duration_seconds / max_iterations) serializing to camelCase maxCost / maxDurationSeconds / maxIterations, dropping None fields. - AgentRunParams accepts an optional budget (Budget instance or dict, snake_case or camelCase). - build_run_payload normalizes the budget and sets executionParams.budget only when non-empty; absence leaves the payload unchanged. - Export Budget from aixplain.v2. - Unit tests for serialization and payload behavior. * feat(v2): deprecate max_iterations in favor of Budget Make Budget the single source of truth for run governance in the v2 SDK. Budget gains a second role as a persistable default on the agent, and every standalone max_iterations surface is deprecated via map + DeprecationWarning. - Persist budget on Agent: new budget field, coerced from dict/Budget in __post_init__, serialized into the create payload as camelCase budget. - Deprecate persisted Agent.max_iterations: warn, fold into budget.max_iterations (Budget wins on conflict), stop emitting standalone maxIterations in the save payload. - Deprecate run-time max_iterations exec param: warn, fold into executionParams.budget.maxIterations, stop emitting standalone executionParams.maxIterations (removed from build_run_payload defaults). - Tests: persisted budget serialization, both deprecation fold paths, conflict precedence, and absence of standalone maxIterations. Thin SDK: no effective-budget merge/precedence (backend owns that). v1 untouched. No session ExecutionConfig surface exists on this branch. * chore: ignore worktree tooling artifacts (#970) * fix(v2): silence max_iterations deprecation on agent load; fix run-path warning stacklevel Review fixes for the max_iterations -> Budget deprecation PR: - Agent.from_dict no longer emits the max_iterations DeprecationWarning when loading backend agents that still carry a legacy top-level maxIterations. @dataclass_json clobbers any in-class from_dict, so the injected decoder is wrapped after decoration: a legacy maxIterations is folded silently into budget.maxIterations (Budget wins on conflict) and the standalone key removed before decoding, so __post_init__ sees max_iterations=None and never warns on load. The explicit Agent(max_iterations=...) constructor still warns. - Run-path deprecation warning (build_run_payload) stacklevel 2 -> 5 and the conflict warning (_fold_iter_into_budget) stacklevel 3 -> 6 so both resolve to the user's agent.run(...) call site rather than SDK internals. - Tests: from_dict with maxIterations emits no warning and folds into budget, budget wins silently on conflict, caller dict not mutated, nested fields still deserialize; constructor still warns; run-path warnings point at the call site. Note: the requested session.py / ExecutionConfig fix does not apply -- no such module or class exists in this codebase; all execution-param logic lives in agent.py and is already consistent. * fix(v2): apply Budget deprecation to Session ExecutionConfig Mirror the agent run path on sessions: ExecutionConfig gains a budget field, max_iterations is removed from EXECUTION_PARAMS_MAP and folded into executionParams.budget.maxIterations (Budget wins on conflict) with a DeprecationWarning. Loading a backend session never warns (legacy maxIterations is folded silently in ExecutionConfig.from_dict / Session.from_dict). Updates stale tests and adds TestExecutionConfigBudget. * refactor(v2): fix session conflict-warning location; align max_iterations handling _fold_iter_into_budget is now pure (returns (normalized, conflicted)); each caller emits the budget-wins UserWarning at its own stacklevel, so the session path warning resolves to to_api_dict() instead of overshooting into sys internals. Shared the conflict message via Agent._BUDGET_ITER_CONFLICT_MSG, aligned the run path to remove max_iterations from _EXEC_PARAMS_MAP + dual-pop both spellings (matching session), dropped the dead budget pop in build_save_payload, and added a session conflict-warning call-site test. * refactor(v2): replace Budget param with agent.budget attribute access --------- Co-authored-by: Abdelrahman Elsheikh <abdelrahman.el-sheikh@aixplain.com>

* feat(v2): redesign Inspector retrieval via marketplace get/search Replace the PrebuiltInspector factory-classmethod API with marketplace-style retrieval: aix.Inspector.get("aixplain/prompt-injection-guard") and aix.Inspector.search(...). Guards are ordinary marketplace models under the guardrails Function, so a fetched guard and a hand-built Inspector are the same type and slot identically into aix.Agent(inspectors=[...]). - Inspector now inherits BaseResource + Get/SearchResourceMixin; .get/.search hit v2/models pinned to function=guardrails and adapt each guard model into a fully-configured Inspector (evaluator -> guard model, tuned action/targets). - Hard-remove PrebuiltInspector, _PREBUILT_REGISTRY, the *_GUARDRAIL_ASSET_ID constants, is_prebuilt_inspector, and the split prebuilt/custom serialization path in agent.py (single Inspector deserialization now). - Add GUARDRAILS to the v2 Function enum. - Tuned defaults resolve by canonical path slug with a safe abort/input fallback, so new guards ship with zero SDK changes; assetInfo is flattened so search results stay REPL-browsable and id-based retrieval still picks the tuned config from the payload path. * fix(v2): correct guard tuned-config resolution by asset name (#950) Verified against production guard models, two issues surfaced: - Tuned defaults never applied. Real guard paths are host/asset-name/instance (e.g. aws/sensitive-information-guardrail/aws), so keying config off the trailing path segment yielded the host ("aws") and every guard fell back to abort/input. In particular the PII/sensitive-information guard wrongly got abort instead of edit (and no editor). Now resolve config from any path segment / assetInfo.assetName, and key _GUARD_CONFIG_DEFAULTS on the real marketplace asset names (detect-prompt-attacks-guardrail, sensitive-information-guardrail, contextual-grounding-check-guardrail). - Replace fictional aixplain/* example paths with the real aws/* paths. Also document that Inspector.search currently returns an empty page: the v2/models/paginate functions filter is a backend no-op (returns 0 for every function, verified with text-generation), tracked as a backend bug. get-by-path is unaffected and verified end-to-end for all three guards. Updates unit tests to the real asset names/paths; adds regression tests for asset-name (middle-segment) resolution. * fix(v2): make BaseResource.context serializable so guards attach to agents (#950) Saving a team agent with a prebuilt guard (aix.Inspector.get(...)) raised "'Inspector' object has no attribute 'context'". Inspector is now a BaseResource, whose `context` field is init=False with no default. An inspector rehydrated via Inspector.from_dict (e.g. nested in an Agent and re-deserialized on save) never gets a bound context, so the attribute is simply absent — dataclasses_json's _asdict calls getattr(obj, "context") before applying the field's `exclude`, raising AttributeError mid-serialization. Give `context` a None default so the attribute always exists; it stays excluded from serialized output and overridable via setattr (get/search/Aixplain binding). Verified end-to-end against the backend: a fetched guard now saves, persists as an inspector, and runs as an input-stage inspector in a team agent. Adds a regression test (Agent carrying a from_dict inspector serializes without leaking context). * test(v2): add guard-by-path + team-agent functional tests and slug unit tests (#950) Functional (verified against test-platform): - get each prebuilt guard by its canonical aws/* path and assert the tuned config resolves correctly (PII -> edit with editor, grounding -> rerun/output, prompt-attacks -> abort/input). - attach a fetched guard to a team agent, save, and run; assert it persists as an Inspector and executes as an inspector step. Unit: - direct coverage for _guard_slugs / _resolve_guard_defaults: asset-name (middle) segment selects config, trailing host segment is ignored, bare id falls back, first matching candidate wins. * refactor(v2): simplify Inspector API — metric judge + string config --------- Co-authored-by: Abdelrahman Elsheikh <abdelrahman.el-sheikh@aixplain.com>

* feat: add v2 Skill asset (knowledge file bundle) Add `aix.Skill`, a managed (non-runnable) asset backed by the /sdk/skill backend endpoints. A Skill is a curated bundle of files and folders that can be attached to an Agent to give it reference material. - Skill: standard lifecycle (save/get/search/delete/clone) via the existing mixins; not runnable (no RunnableResourceMixin), matching the backend. - SkillNode + Skill.File()/Skill.Folder() constructors; files accept a remote url or a local path (auto-uploaded via FileUploader). - File-tree reconciliation behind save(save_subcomponents=True) (BETA): diffs children vs. last-synced state and issues the node create/update/delete calls, so the per-node endpoints stay hidden from developers. - download(to=...), get_node(path_or_id), refresh(). - Agent.skills: List[Union[str, Skill]], passed like tools/agents; resolves to ids, participates in save(save_subcomponents=True), and validates that attached skills are saved first. - Wired into core.py (aix.Skill) and v2 __init__ exports (Skill, SkillNode). * feat: attach skills to agents using the tools wire design The backend rejects bare skill id strings on agent create ("skills.each value ... must be either object or array"). Skills must be sent as objects, exactly like tools — not like sub-agents. - Skill is now a ToolableMixin and implements as_tool() -> {id, name, description, supplier, type: "skill", version, asset_id}. - Agent.build_save_payload serializes `skills` through the same as_tool() -> _normalize_tool_dict_for_api path as `tools`, accepting Skill instances, dicts, or id strings. Verified end-to-end against DEV: agent create accepts the skill and the agent uses it at runtime. (Round-trip read is pending a backend change to echo `skills` in the agent GET response.) * fix: upload skill files via the temp upload-url endpoint The permanent pre-signed-URL endpoint (/sdk/file/upload-url) 500s on DEV, while the temp endpoint (/sdk/file/upload/temp-url) works. Match file.py's Resource default (is_temp=True) so local-path skill files upload and attach. * refactor: folder-based Skill authoring; hide node details Adopt the Claude-style skill design. A skill is authored from a local folder and the file tree is uploaded/managed internally — the per-node API is no longer part of the developer surface. - aix.Skill(folder="./skills/pdf-filler"): parse SKILL.md YAML frontmatter (name, description, requires -> required_tools) + markdown body (instructions); the folder is uploaded on save(). - save() creates the asset, then walks the folder and builds the backend file/folder tree internally (no save_subcomponents flag, no SkillNode). - Remove public node surface: SkillNode, Skill.File/Folder, children, get_node, save_subcomponents (+ drop SkillNode export). - Keep get/search/download/clone/delete/refresh/as_tool; expose read-only name/description/tags/privacy/required_tools/instructions. Verified on DEV: folder authoring + frontmatter parse + bundle upload + download + search + clone + agent attachment all work. * refactor(v2): rename Skill authoring param to file_path; accept single .md files --------- Co-authored-by: Abdelrahman Elsheikh <abdelrahman.el-sheikh@aixplain.com>

Co-authored-by: aix-ahmet <ahmet.gunduz@aixplain.com>

* feat(v2): unify session attachments param; support audio-as-prompt * feat(v2): support attachments on non-session agent.run * feat(v2): auto-detect attachment type from url/path extension * Update attachments parameter documentation Clarified the description of the attachments parameter. --------- Co-authored-by: aix-ahmet <ahmet.gunduz@aixplain.com>

…Result (#983)

The v2 SDK dropped reasoning-model chain-of-thought text from model run responses. Add a reasoning_content field to both Message (non-streaming) and StreamChunk (streaming) and carry it through the OpenAI-style stream chunk parsing path. Closes #985

…ion=) (#984)

…1) (#991)

… discovery (#989)

# Conflicts: # aixplain/v2/inspector.py # tests/unit/v2/test_inspector.py

…collisions (#993) Duplicate test basenames (test_model.py, test_tool.py, test_session.py, test_trigger.py, test_api_key.py, test_rlm.py) exist in both tests/unit/v2 and tests/functional/v2 (and tests/functional/model). Without __init__.py files, pytest's default prepend import mode maps them to the same module name and aborts collection with 'import file mismatch'. --import-mode=importlib gives each file a fully-qualified module name, so unique basenames are no longer required.

aix-ahmet and others added 30 commits June 10, 2026 09:40

fix(v2): add governance to AgentResponseData (#957)

32b8b39

fixed request id mapping (#964)

171c82a

added warning for saving agent with integration (#965)

db5dbad

* added warning for saving agent with integration * added name comment for clarification

wrapped non-json (#968)

0cc1e52

* wrapped non-json * removed debug

Added user id to headers (#969)

d03ce21

fix: Revert "wrapped non-json (#968)" (#974)

1282bf8

This reverts commit 0cc1e52.

ENG-3191:check url before polling (#975)

9ea569b

* ENG-3191:check url before polling * ENG-3191: fix id in test

final shape doesn't include url (#977)

1b87061

* final shape doesn't include url * ensure same behavior on v2

parse data error (#979)

1f42985

test(v2): functional regression tests for agent LLM persistence acros…

f458e4a

…s fetch/save (#982) Pins the 0.2.43/0.2.44 regression where Agent.get() fell back to DEFAULT_LLM and any save() overwrote the agent's configured LLM.

feat(v2): object-based tool parameter overrides (#981)

286a011

Co-authored-by: aix-ahmet <ahmet.gunduz@aixplain.com>

fix(v2): surface diagnostic error codes nested under data in AgentRun…

d188381

…Result (#983)

feat(v2): manage sessions via aix.Session; polymorphic agent.run(sess…

5ef8c33

…ion=) (#984)

Merge branch 'test' into development

7d4f2df

feat(inspector): emit presetId for fetched marketplace guards (BUG-76…

eab52d0

…1) (#991)

feat(v2): first-class aix.Trigger asset with time schedules and event…

4daee95

… discovery (#989)

Merge remote-tracking branch 'origin/test' into development

c01ae56

# Conflicts: # aixplain/v2/inspector.py # tests/unit/v2/test_inspector.py

aix-ahmet and others added 3 commits July 21, 2026 11:50

Merge remote-tracking branch 'origin/test' into development

4a0e833

Merge remote-tracking branch 'origin/test' into development

448ad66

test: prune flaky/failing functional tests and trim CI matrix (#1003)

04dea96

hadi-aix merged commit 43f20a4 into test Jul 23, 2026
1 check passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Merge to test 20260723#1004

Merge to test 20260723#1004
hadi-aix merged 33 commits into
testfrom
development

hadi-aix commented Jul 23, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

Uh oh!

Conversation

hadi-aix commented Jul 23, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants