Add ollama llm by ChiefGyk3D · Pull Request #67 · ChiefGyk3D/penguin-overlord

ChiefGyk3D · 2026-01-21T03:58:08Z

No description provided.

github-actions · 2026-01-21T03:58:20Z

Dependency Review

The following issues were found:

✅ 0 vulnerable package(s)
✅ 0 package(s) with incompatible licenses
✅ 0 package(s) with invalid SPDX license definitions
⚠️ 1 package(s) with unknown licenses.

See the Details below.

Snapshot Warnings

⚠️: No snapshots were found for the head SHA 48d6b09.

Ensure that dependencies are being submitted on PR branches and consider enabling retry-on-snapshot-warnings. See the documentation for more information and troubleshooting advice.

License Issues

.github/workflows/dependency-scan.yml

Package	Version	License	Issue Type
actions/upload-artifact	6..	Null	Unknown License

OpenSSF Scorecard

Package

Version

Score

Details

actions/actions/upload-artifact

6.*.*

🟢 5.7

Details

Check	Score	Reason
Code-Review	🟢 10	all changesets reviewed
Packaging	⚠️ -1	packaging workflow not detected
Maintained	🟢 5	4 commit(s) and 2 issue activity found in the last 90 days -- score normalized to 5
Binary-Artifacts	🟢 10	no binaries found in the repo
Dangerous-Workflow	🟢 10	no dangerous workflow patterns detected
CII-Best-Practices	⚠️ 0	no effort to earn an OpenSSF best practices badge detected
Token-Permissions	⚠️ 0	detected GitHub workflow tokens with excessive permissions
Pinned-Dependencies	⚠️ 1	dependency not pinned by hash detected -- score normalized to 1
Fuzzing	⚠️ 0	project is not fuzzed
License	🟢 10	license file detected
Signed-Releases	⚠️ -1	no releases found
Security-Policy	🟢 9	security policy file detected
Branch-Protection	⚠️ 0	branch protection not enabled on development/release branches
SAST	🟢 10	SAST tool is run on all commits

pip/aiohttp

3.13.3

Unknown

pip/google-genai

1.14.0

Unknown

pip/matplotlib

3.10.8

Unknown

Scanned Files

.github/workflows/dependency-scan.yml
requirements.txt

🤖 New Features: - OllamaClient utility for local LLM integration - Smart Arch Linux roasts using LLM (optional, env-controlled) - Contextual, personalized responses based on message content - Graceful fallback to classic regex jokes if Ollama unavailable 🎯 Arch Banter Enhancements: - ARCH_BANTER_LLM env var enables LLM mode - Generates contextual roasts from actual user messages - Falls back to 90+ classic jokes automatically - Maintains cooldowns and statistics 📚 Infrastructure: - penguin-overlord/utils/ollama_client.py - Async LLM client - Configurable via OLLAMA_MODEL, OLLAMA_ENABLED env vars - 8-second timeout with async processing (non-blocking) - Supports gemma2:2b, llama3.2, llama3.1 models 🔧 Configuration: - requirements.txt: Added ollama==0.6.1 - Comprehensive docs/OLLAMA_LLM.md with setup guide - Default: LLM disabled, opt-in feature - Privacy-first: 100% local processing, no cloud APIs 💡 Future Ready: - Quote search with semantic understanding - Contextual quote insights/explanations - Smart recommendations - Conversational trivia Tested with gemma2:2b model. Falls back gracefully if Ollama not installed/running. Zero breaking changes - fully backward compatible.

📦 Dependency Updates: - boto3: 1.41.2 → 1.42.30 - aiohttp: 3.13.2 → 3.13.3 (security fix) - matplotlib: 3.10.7 → 3.10.8 - GitHub Actions upload-artifact: v5 → v6 🤖 Robust Multi-Model LLM Support: - Added MODEL_CONFIGS with optimal settings for 17+ models - Gemma 2/3 (2b, 4b, 9b, 12b) - Google's efficient models - Llama 3.x (1b, 3b, 8b, 70b) - Meta's powerful models - Phi 3/4 (mini, medium, latest) - Microsoft's compact models - Qwen 2.5 (0.5b, 3b, 7b) - Alibaba's multilingual models - Mistral/Mixtral (7b, 8x7b) - High-quality open models 🎯 Smart Configuration: - Model-specific temperature, token limits, top_k values - Context window sizes (8K to 128K tokens) - Automatic fallback for unknown models - Partial matching (gemma3:4b-instruct matches gemma3:4b) - Family matching for variants All changes maintain backward compatibility

Changes to ollama_client.py: - Support configurable OLLAMA_HOST and OLLAMA_PORT env vars (previously hardcoded to localhost, couldn't use remote server) - Use ollama.Client(host=...) instead of module-level ollama.generate() for proper remote server support - Fix asyncio.get_event_loop() deprecation (Python 3.10+) by using asyncio.get_running_loop() instead - Add qwen3:14b to MODEL_CONFIGS - Smart URL construction: auto-prepend http://, handle port in URL Changes to .env.example: - Document all Ollama configuration variables: OLLAMA_ENABLED, OLLAMA_HOST, OLLAMA_PORT, OLLAMA_MODEL, ARCH_BANTER_LLM Tested with Ollama server at 192.168.214.10:11434.

…ncement

+
+EMOJI_PATTERN = re.compile(
+    "["
+    "\U0001F600-\U0001F64F"  # emoticons


+
+EMOJI_PATTERN = re.compile(
+    "["
+    "\U0001F600-\U0001F64F"  # emoticons


+
+EMOJI_PATTERN = re.compile(
+    "["
+    "\U0001F600-\U0001F64F"  # emoticons


+
+EMOJI_PATTERN = re.compile(
+    "["
+    "\U0001F600-\U0001F64F"  # emoticons


+
+EMOJI_PATTERN = re.compile(
+    "["
+    "\U0001F600-\U0001F64F"  # emoticons


+    from ai.features.moderation import (
+        ModerationCategory, ModerationResult,
+        ACTIONS_REQUIRING_REVIEW, ACTION_SEVERITY,
+    )


In general, to fix an unused-import problem you either (a) remove the imported names from the import statement, or (b) if none of the imported names are used, remove the entire import block. If there is a fallback definition for the same unused names (as here), that fallback code can also be removed to keep the file consistent and avoid dead code.

In this file, the best targeted change is to stop importing ACTIONS_REQUIRING_REVIEW and ACTION_SEVERITY from ai.features.moderation / penguin_overlord.ai.features.moderation, and to remove the corresponding default assignments in the deepest except ImportError block, while leaving the rest of the imports and logic intact. Concretely:

In the first try block at lines 64–67, keep importing ModerationCategory and ModerationResult, but drop ACTIONS_REQUIRING_REVIEW and ACTION_SEVERITY from the import list.

In the inner try block at lines 70–73, do the same.

In the final except ImportError block at lines 75–76, remove the assignments to ACTIONS_REQUIRING_REVIEW and ACTION_SEVERITY, because those names are no longer needed.

No new methods, imports, or definitions are required to implement this; we are only narrowing the imported symbols and deleting the now-unnecessary fallbacks.

+import logging
+import os
+from datetime import datetime, timedelta, timezone
+from typing import Optional, List


To fix the problem, we should remove the unused symbol Optional from the typing import while keeping List, which is used in the return annotation of _get_env_list. This removes an unnecessary dependency and quiets the static analysis warning without changing runtime behavior.

Concretely, in penguin-overlord/cogs/ai_moderation.py, adjust the import on line 32 from from typing import Optional, List to from typing import List. No other code changes are required, and no additional methods or imports are needed.

+from .ollama_provider import OllamaProvider
+from .gemini_provider import GeminiProvider
+from .queue import RequestQueue
+from .guardrails import Guardrails, GuardrailConfig, GuardrailResult, FEATURE_GUARDRAIL_DEFAULTS


To fix the issue, remove the unused GuardrailResult name from the import statement while keeping the other imported names intact. This eliminates the unnecessary dependency on that symbol without altering functionality, since it is not referenced anywhere in the module.

Concretely, in penguin-overlord/ai/manager.py, locate the line:

from .guardrails import Guardrails, GuardrailConfig, GuardrailResult, FEATURE_GUARDRAIL_DEFAULTS

and edit it to:

from .guardrails import Guardrails, GuardrailConfig, FEATURE_GUARDRAIL_DEFAULTS

No other code changes or additional imports are needed.

+from collections import deque
+from dataclasses import dataclass, field
+from enum import Enum
+from typing import Dict, List, Optional, Set, Tuple


To fix the problem, remove the unused symbol Set from the typing import so that only the actually used types remain imported. This preserves existing functionality and behavior while cleaning up the dependency surface and satisfying the static analysis rule.

Concretely, in penguin-overlord/ai/guardrails.py, locate the line:

from typing import Dict, List, Optional, Set, Tuple

and edit it to remove Set from the imported names:

from typing import Dict, List, Optional, Tuple

No other code changes are required, and no new methods, imports, or definitions are needed.

+
+import logging
+import re
+import time


To fix an unused import, you remove the import statement for the module that is not referenced anywhere in the file. This does not change runtime behavior if the module truly isn’t used, but it cleans up the dependency graph and the code.

In this file, the single best fix is to delete the line import time at line 19 in penguin-overlord/ai/guardrails.py, leaving all other imports intact. No additional methods, definitions, or imports are required; we are only removing the unused import. Ensure that only this line is removed and that the surrounding imports remain unchanged.

ChiefGyk3D added 2 commits February 7, 2026 10:19

ChiefGyk3D force-pushed the add-ollama-llm branch from 335b506 to be1ce7f Compare February 7, 2026 15:21

ChiefGyk3D added 2 commits February 8, 2026 20:55

AI Integrations to tons of features, preparing for AI Moderation enha…

bbe916b

…ncement

github-advanced-security AI found potential problems Feb 14, 2026

View reviewed changes

codeql fixes and aiosqlite for database for moderation and such

8bf740b

github-advanced-security AI found potential problems Feb 14, 2026

View reviewed changes

Add Solana (SOL) to cryptocurrency tips

48d6b09

@@ -62,20 +62,20 @@
             # Try to import moderation types
             try:
                 from ai.features.moderation import (
-                    ModerationCategory, ModerationResult,
-                    ACTIONS_REQUIRING_REVIEW, ACTION_SEVERITY,
+                    ModerationCategory,
+                    ModerationResult,
                 )
             except ImportError:
                 try:
                     from penguin_overlord.ai.features.moderation import (
-                        ModerationCategory, ModerationResult,
-                        ACTIONS_REQUIRING_REVIEW, ACTION_SEVERITY,
+                        ModerationCategory,
+                        ModerationResult,
                     )
                 except ImportError:
-                    ACTIONS_REQUIRING_REVIEW = {'kick', 'ban'}
-                    ACTION_SEVERITY = {}
+                    pass
             def _get_env_list(key: str) -> List[str]:
                 """Get a comma-separated env var as a list of strings."""
                 val = os.getenv(key, '').strip()

Conversation

ChiefGyk3D commented Jan 21, 2026

Uh oh!

github-actions Bot commented Jan 21, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Dependency Review

Snapshot Warnings

License Issues

.github/workflows/dependency-scan.yml

OpenSSF Scorecard

Scanned Files

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Check warning

Check warning

Check warning

Check warning

Check warning

Check notice

Copilot Autofix

Check notice

Copilot Autofix

Check notice

Copilot Autofix

Check notice

Copilot Autofix

Check notice

Copilot Autofix

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

github-actions Bot commented Jan 21, 2026 •

edited

Loading