feat(frontend): add frontend#7
Merged
Merged
Conversation
📝 WalkthroughWalkthroughThis PR adds a complete Vue 3 frontend application alongside backend API expansions to enable a fully functional chat and knowledge-management interface for the Mind-Bus AI Agent system. The backend gains new endpoints for thread management, document operations, memory queries, and tool control. The frontend provides authentication, chat interface, document upload/management, memory browsing, and tool control views, supported by comprehensive architecture and integration documentation. ChangesFrontend Application and API Expansion
🎯 4 (Complex) | ⏱️ ~60 minutes 🚥 Pre-merge checks | ✅ 3 | ❌ 2❌ Failed checks (1 warning, 1 inconclusive)
✅ Passed checks (3 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing Touches📝 Generate docstrings
🧪 Generate unit tests (beta)
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
Review Summary by QodoAdd complete frontend implementation with Vue 3 and backend API enhancements
WalkthroughsDescription• **Complete frontend implementation** with Vue 3 components for chat, documents, tools, and memory management • **Backend API enhancements** including tool management endpoints, complete document CRUD operations, memory deletion, and chat thread listing • **Authentication system** with JWT-based login, Pinia store integration, and route guards • **API client setup** with axios interceptors for automatic token attachment and error handling • **Build and deployment configuration** including Vite setup, Docker containerization, and docker-compose integration • **Comprehensive documentation** covering integration guide, agent architecture, technical skills reference, and quick start guide • **Frontend state management** with Pinia stores for authentication and chat functionality • **GitHub agent configuration** for project-specific AI assistant Diagramflowchart LR
A["Backend APIs<br/>tools, documents,<br/>memory, chat"] -- "HTTP/REST" --> B["Axios Client<br/>with Auth<br/>Interceptors"]
B -- "JWT Token" --> C["Vue 3<br/>Frontend<br/>Components"]
C -- "State" --> D["Pinia Stores<br/>auth, chat"]
E["Router with<br/>Auth Guards"] -- "Navigation" --> C
F["Docker Compose<br/>Frontend Service"] -- "Builds & Serves" --> C
G["Documentation<br/>Integration, Agent,<br/>Skills, Quickstart"] -- "References" --> A
G -- "References" --> C
File Changes1. apps/api/routes/tools.py
|
Code Review by Qodo
1.
|
![qodo-code-review[bot]](https://avatars.githubusercontent.com/in/484649?s=60&v=4){kind=small}
- Introduced a comprehensive skill for building Vue 3 components, including state management with Pinia, routing, form handling, API integration, and SCSS styling. - Added a memory system skill detailing episodic, semantic, and correction memory types, along with workflows for memory extraction, storage, and retrieval. - Implemented a RAG retrieval skill that covers document chunking, hybrid search strategies, and caching mechanisms for optimized retrieval performance.
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 19
🧹 Nitpick comments (2)
apps/api/routes/documents.py (1)
10-12: 🏗️ Heavy liftIn-memory store + global counter won't survive restarts or multi-worker deploys.
_documents_storeand_document_counterare process-local, so uploaded docs vanish on restart and are invisible to other API workers;doc_<n>ids also collide across replicas. Themigrations/0001_initial_schema.sqldocumentstable already exists for this. Acknowledged by your TODO comment — flagging so it isn't shipped to production silently.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@apps/api/routes/documents.py` around lines 10 - 12, The in-memory _documents_store and _document_counter are process-local and will not survive restarts or multiple workers; replace their usage with persistent DB-backed operations against the existing documents table (use the migration-created table) by implementing CRUD functions that insert/select/update/delete rows instead of mutating _documents_store/_document_counter, update any handlers referencing _documents_store or _document_counter to call those DB functions (e.g., create_document, get_document, list_documents, delete_document), and ensure generated IDs use the DB primary key (or UUIDs) to avoid replica collisions and survive restarts.apps/api/routes/tools.py (1)
66-96: ⚖️ Poor tradeoff
_tools_stateis process-local — toggles won't persist or propagate across workers.
PATCH /tools/{tool_id}mutates an in-memory dict, so enable/disable state is lost on restart and inconsistent across API replicas (one worker may allow a tool the next blocks). Acceptable for a demo, but flagging before production.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@apps/api/routes/tools.py` around lines 66 - 96, The current implementation uses a process-local dict _tools_state mutated by toggle_tool and read by list_tools, so toggles don't persist or propagate across workers; replace this with a shared persistent store (e.g., a database table or Redis) and update list_tools and toggle_tool to read/write from that store instead of _tools_state: keep AVAILABLE_TOOLS as the canonical tool list, implement a persistence layer (e.g., get_tool_state(tool_id), set_tool_state(tool_id, enabled)) and call those from list_tools and toggle_tool (and remove reliance on _tools_state) so state is durable and consistent across replicas and restarts.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Inline comments:
In @.agents/skills/mindbus-skill/SKILL.md:
- Around line 3-8: The SKILL.md currently contains template placeholders;
replace them with concrete, actionable documentation for the mindbus-skill: add
a clear "When to use" section, a concise "What it does" summary, and a
step-by-step "How to use" that maps to the Mind-Bus flow and state machine steps
(Retrieve → Plan → Execute → Respond → Reflect → Compress → Return); include
examples of inputs/outputs, expected agent behaviors, failure modes, and sample
prompts or API calls showing each state transition so agents can follow the
Retrieve/Plan/Execute/Respond/Reflect/Compress/Return sequence reliably when
invoking mindbus-skill.
In `@apps/api/routes/documents.py`:
- Around line 74-79: The Content-Disposition header is vulnerable because
doc['filename'] is interpolated raw; update the header assembly in the
StreamingResponse return so the filename is sanitized and properly quoted:
create a small sanitizer that strips control characters (CR/LF and other ASCII
<0x20 and 0x7f), removes or escapes quotes/semicolons/backslashes, and then use
a quoted filename and an RFC5987-encoded filename* (e.g., percent-encode UTF-8)
when building the headers instead of f"attachment; filename={doc['filename']}"
so the StreamingResponse headers include a safe quoted filename and a filename*
fallback; locate and modify the header creation around StreamingResponse and
doc['filename'] to apply this sanitizer/encoding.
- Around line 25-48: The handler currently decodes uploaded bytes to UTF-8
(contents.decode(...)) which corrupts binary files; change the store to keep the
raw bytes by assigning the bytes object (contents) into
_documents_store[doc_id]["content"] and keep "size" as len(contents); remove the
.decode(...) usage and only decode when you explicitly need text. Also update
the corresponding download_document logic to return the stored bytes with the
correct media_type instead of re-encoding a lossy string. Finally, change the
exception re-raise to chain the original exception (raise HTTPException(...)
from e) so the original error is preserved in logs; use the same identifiers
_documents_store, doc_id, download_document, and the existing
logger/HTTPException in your edits.
In `@apps/api/routes/memory.py`:
- Around line 196-200: The route currently converts memory_id with
UUID(memory_id) and wraps everything in a blanket except, causing malformed IDs
(ValueError) and "not found" deletes to become 500; update the handler to first
validate the id by catching ValueError from UUID(memory_id) and raise
HTTPException(status_code=400, detail="Invalid memory_id"), then call
memory_manager.delete_memory(UUID(memory_id), user_id) and handle the delete
result or specific not-found exception (e.g., KeyError or a custom NotFoundError
thrown by memory_manager.delete_memory) by raising
HTTPException(status_code=404, detail="Memory not found"); only catch and
rethrow other unexpected exceptions as 500. Ensure you reference UUID,
ValueError, memory_manager.delete_memory, and HTTPException when making these
changes.
- Around line 165-192: The router is mounted without a prefix so `@router.get`("")
and `@router.delete`("/{memory_id}") become GET / and DELETE /{memory_id}; add a
"/memory" prefix (either change APIRouter() to APIRouter(prefix="/memory") in
the module that defines router or include_router(memory.router,
prefix="/memory") where the router is mounted) so get_memories_alias remains GET
/memory and delete_memory becomes DELETE /memory/{memory_id}; also implement the
missing delete_memory method on LongTermMemoryManager in memory/long_term.py
(matching the signature used by the handler) or change the handler to call the
existing removal method name—ensure the handler calls a real
LongTermMemoryManager API (referenced symbols: APIRouter, get_memories_alias,
delete_memory handler, memory_manager.delete_memory, LongTermMemoryManager).
In `@apps/api/routes/tools.py`:
- Around line 103-108: The run_tool_endpoint currently checks _tools_state
(populated from AVAILABLE_TOOLS) and returns 403 for names not present there,
causing registered tools in TOOL_REGISTRY (e.g., code_exec) to be blocked and
allowing non-registered IDs (file_read/file_write/api_call) to pass but later
fail in run_tool; fix by aligning AVAILABLE_TOOLS with TOOL_REGISTRY at startup
(or derive AVAILABLE_TOOLS from TOOL_REGISTRY), update _tools_state to reflect
TOOL_REGISTRY entries, and change error handling in run_tool_endpoint/run_tool
to return 404 for unknown tool ids while keeping 403 for explicitly disabled
tools so that TOOL_REGISTRY, AVAILABLE_TOOLS, _tools_state, run_tool_endpoint
and run_tool remain consistent.
In `@docs/INTEGRATION.md`:
- Around line 434-438: The docs currently show an incorrect startup sequence
using "cd /" before running "python -m apps.api.main"; update the snippet to
either remove the cd line and instruct "run from repository root" or replace "cd
/" with a repository-root placeholder (e.g., "cd <repo-root>" or the actual repo
directory) so "python -m apps.api.main" is executed from the project context
rather than the filesystem root.
In `@docs/QUICKSTART.md`:
- Line 77: The PostgreSQL command in the docs uses a different database name
("ai_agent") than the rest of the project ("mindbus"); update the docker-compose
command shown (the line starting with "docker-compose -f
deploy/docker-compose.yml exec postgres psql -U postgres -d ai_agent") to use
"-d mindbus" so the README/QUICKSTART is consistent with the project's canonical
DB name.
In `@frontend/Dockerfile`:
- Around line 15-25: The Dockerfile runs the final CMD ("serve" invocation) as
root; switch to the non-root node user by adding a USER directive (e.g., USER
node) in the production stage before CMD and ensure the /app and /app/dist files
are owned/readable by that user (adjust COPY or run chown during the build or
set permissions) so the serve command can run without root; locate the
production stage that uses CMD ["serve", "-s", "dist", "-l", "5173"] and add the
USER change and any necessary ownership/permission steps to that stage.
- Around line 6-8: The Dockerfile uses the COPY pattern "COPY package*.json ./"
followed by "RUN npm ci" but frontend/package-lock.json is not present/tracked,
causing npm ci to fail; fix this by either adding and committing
frontend/package-lock.json to the repo so it is available during build or change
the Dockerfile to not rely on npm ci (e.g., copy only package.json and use npm
install) — update the COPY/CMD usage around the COPY package*.json ./ and RUN
npm ci entries accordingly so the build no longer requires a missing
package-lock.json.
In `@frontend/src/services/api.ts`:
- Around line 19-23: The interceptor currently treats any 401 as an auth failure
and calls useAuthStore().logout() then forces window.location.href = '/login',
which also triggers on failed /auth/token login attempts and prevents the
LoginView catch from showing "Invalid credentials"; update the interceptor so it
ignores 401s coming from the login endpoint (e.g. requests whose URL endsWith
'/auth/token' or matches the auth route) and do not redirect when the app is
already on '/login' (check window.location.pathname) — only call
authStore.logout() and navigate to login for other 401s; refer to the existing
useAuthStore(), authStore.logout(), and the window.location.href redirect in the
interceptor to locate where to change logic.
In `@frontend/src/stores/auth.ts`:
- Around line 19-20: You’re persisting the JWT in localStorage (token.value =
response.data.access_token; localStorage.setItem('token', ...)), which exposes
it to XSS; instead, change the client to stop storing the access token in
localStorage and coordinate with the backend to set the token as an HttpOnly,
Secure cookie (SameSite as appropriate) on authentication responses; remove the
client-side localStorage.setItem call and update the request flow (and the
interceptor in api.ts that currently attaches the Authorization header) to rely
on cookie-based auth so the browser sends the cookie automatically, and ensure
any client code referencing token.value is adjusted to read auth state from a
non-secret source (e.g., isAuthenticated flag) rather than raw token.
- Around line 12-29: The store currently only decodes the JWT inside login(),
leaving user null after page refresh; add a Base64URL-safe decoder function
(e.g., decodeUser or decodeJWTPayload) that: converts '-' to '+', '_' to '/',
adds '=' padding as needed, then JSON.parse(atob(...)) inside a try/catch and
returns the parsed payload or null on error. Call this decoder during store
initialization to populate user from token (the ref token created from
localStorage.getItem('token')) and also use it after login() to set user.value;
ensure any decode errors set user.value = null instead of throwing.
In `@frontend/src/stores/chat.ts`:
- Around line 27-68: Each action (fetchThreads, fetchMessages, sendMessage)
fails to clear prior errors so error.value remains set after subsequent
successes; at the start of each function reset error.value = '' (or null) before
setting loading.value = true so the UI error banner clears when a new request
begins; update fetchThreads, fetchMessages, and sendMessage to reset error.value
at their entry.
In `@frontend/src/style.scss`:
- Line 104: Remove the stray HTML closing tag "</style>" from the SCSS file
(frontend/src/style.scss) — this tag is invalid in .scss and breaks the build;
locate the literal "</style>" in the file (around where the stylesheet rules
end) and delete that token so the file contains only valid SCSS/CSS content.
In `@frontend/src/views/DocumentsView.vue`:
- Around line 130-144: The downloadDocument function currently sets
link.download = '' and never revokes the object URL; update downloadDocument to
extract the filename from response.headers['content-disposition'] (parse the
filename* or filename parameter) or fall back to `${docId}` with an appropriate
extension, set link.download = filename, append the anchor to document.body,
call link.click(), then revoke the object URL with
window.URL.revokeObjectURL(url) (e.g., inside a setTimeout or after click) and
remove the anchor element to avoid leaks; reference the downloadDocument
function, link.download assignment, and window.URL.createObjectURL usage when
making these changes.
- Around line 117-119: Guard the upload progress handler in onUploadProgress so
it handles a possibly undefined e.total (compute percent only when e.total is a
number, default to 0 or e.loaded when unknown) and assign the result to
uploadProgress.value; after creating a blob URL for download, call
window.URL.revokeObjectURL(url) once the link is clicked or after triggering the
download to avoid leaking object URLs, and set link.download to doc.filename (or
use a parsed filename from Content-Disposition) instead of an empty string so
the original filename is preserved; also review the isDragging state (it’s never
set true) and either remove it or wire it to the drag events to make it
meaningful.
In `@frontend/src/views/MemoryView.vue`:
- Around line 28-33: The empty-state check uses memories.length but the UI
renders filteredMemories, so when all items are filtered out the view is blank;
change the conditional to base the empty-state on filteredMemories (e.g.,
replace the v-else-if that checks memories.length === 0 with a check against
filteredMemories.length === 0) so the <div class="empty-state"> shows whenever
filteredMemories is empty while the grid still iterates over filteredMemories
(ensure the v-for still references filteredMemories and keys on memory.id).
In `@SKILL.md`:
- Around line 43-61: Two fenced code blocks in SKILL.md are missing language
tags (the one containing the project tree that starts with "apps/api/" and the
other similar fenced block later in the file); add explicit language identifiers
(e.g., "text", "bash", or "plaintext") to both opening triple-backtick fences so
markdownlint MD040 stops failing and the blocks render with a language marker.
---
Nitpick comments:
In `@apps/api/routes/documents.py`:
- Around line 10-12: The in-memory _documents_store and _document_counter are
process-local and will not survive restarts or multiple workers; replace their
usage with persistent DB-backed operations against the existing documents table
(use the migration-created table) by implementing CRUD functions that
insert/select/update/delete rows instead of mutating
_documents_store/_document_counter, update any handlers referencing
_documents_store or _document_counter to call those DB functions (e.g.,
create_document, get_document, list_documents, delete_document), and ensure
generated IDs use the DB primary key (or UUIDs) to avoid replica collisions and
survive restarts.
In `@apps/api/routes/tools.py`:
- Around line 66-96: The current implementation uses a process-local dict
_tools_state mutated by toggle_tool and read by list_tools, so toggles don't
persist or propagate across workers; replace this with a shared persistent store
(e.g., a database table or Redis) and update list_tools and toggle_tool to
read/write from that store instead of _tools_state: keep AVAILABLE_TOOLS as the
canonical tool list, implement a persistence layer (e.g.,
get_tool_state(tool_id), set_tool_state(tool_id, enabled)) and call those from
list_tools and toggle_tool (and remove reliance on _tools_state) so state is
durable and consistent across replicas and restarts.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: 7e3ed5e2-0fce-4a6e-a758-d8e432683e9c
📒 Files selected for processing (30)
.agents/skills/mindbus-skill/SKILL.md.github/agents/mind-bus-expert.agent.mdAGENT.mdSKILL.mdapps/api/routes/chat.pyapps/api/routes/documents.pyapps/api/routes/memory.pyapps/api/routes/tools.pydeploy/docker-compose.ymldocs/INTEGRATION.mddocs/QUICKSTART.mdfrontend/.gitignorefrontend/Dockerfilefrontend/README.mdfrontend/index.htmlfrontend/package.jsonfrontend/src/App.vuefrontend/src/main.tsfrontend/src/router/index.tsfrontend/src/services/api.tsfrontend/src/stores/auth.tsfrontend/src/stores/chat.tsfrontend/src/style.scssfrontend/src/views/ChatView.vuefrontend/src/views/DocumentsView.vuefrontend/src/views/LoginView.vuefrontend/src/views/MemoryView.vuefrontend/src/views/ToolsView.vuefrontend/tsconfig.jsonfrontend/vite.config.ts
Comment on lines
+165
to
+192
| @router.get("", response_model=dict) | ||
| async def get_memories_alias( | ||
| memory_type: Optional[str] = None, | ||
| limit: int = 50, | ||
| user: dict = Depends(get_current_user), | ||
| ) -> dict: | ||
| """Alias for /memories endpoint for frontend compatibility""" | ||
| user_id = _require_user_id(user) | ||
| validated_memory_type = _validate_memory_type(memory_type) | ||
| bounded_limit = max(1, min(limit, 200)) | ||
|
|
||
| try: | ||
| memories = memory_manager.retrieve_memories( | ||
| user_id=user_id, | ||
| thread_id=None, | ||
| memory_type=validated_memory_type, | ||
| key_pattern=None, | ||
| limit=bounded_limit, | ||
| ) | ||
| except Exception as exc: | ||
| raise HTTPException(status_code=503, detail="Memory storage unavailable") from exc | ||
|
|
||
| serialized = [_serialize_memory(memory) for memory in memories] | ||
| return {"memories": serialized, "count": len(serialized)} | ||
|
|
||
|
|
||
| @router.delete("/{memory_id}") | ||
| async def delete_memory(memory_id: str, user: dict = Depends(get_current_user)) -> dict: |
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
#!/bin/bash
# Find where the memory router is included and with what prefix
rg -nP --type=py 'include_router\([^)]*memory' -C2
fd -e py | rg -i 'memory' | head
rg -nP --type=py 'APIRouter\(' apps/api/routes/memory.pyRepository: Iro96/Mind-Bus
Length of output: 584
🏁 Script executed:
#!/bin/bash
set -euo pipefail
# 1) How the memory router is mounted
echo "== apps/api/main.py include_router block =="
nl -ba apps/api/main.py | sed -n '1,120p'
echo
echo "== include_router(memory.router) exact line =="
rg -n "include_router\(memory\.router\)" -n apps/api/main.py
echo
# 2) How the APIRouter is defined in apps/api/routes/memory.py
echo "== apps/api/routes/memory.py router definition + relevant endpoints =="
nl -ba apps/api/routes/memory.py | sed -n '1,260p'
# 3) Check for other root-level single-segment path deletes (potential conflicts)
echo
echo "== Other DELETE routes shaped like '/{...}' =="
rg -n --type=py "`@router`\.delete\(\"/\{[^/}]+\}\"\)" apps/api/routes -SRepository: Iro96/Mind-Bus
Length of output: 163
🏁 Script executed:
#!/bin/bash
set -euo pipefail
# Helper: show file with line numbers using awk (no nl dependency)
show_lines() {
local file="$1"
local start="${2:-1}"
local end="${3:-200}"
awk -v s="$start" -v e="$end" 'NR>=s && NR<=e {printf "%6d\t%s\n", NR, $0}' "$file"
}
echo "== apps/api/main.py (first 140 lines) =="
show_lines apps/api/main.py 1 140
echo
echo "== include_router(memory.router) exact match =="
rg -n "include_router\(memory\.router\)" apps/api/main.py || true
echo
echo "== apps/api/routes/memory.py router definition + endpoints (up to ~260) =="
show_lines apps/api/routes/memory.py 1 260
echo
echo "== Other DELETE routes shaped like '/{...}' (any router) =="
rg -n --type=py "`@router`\.delete\(\"/\{[^/}]+\}\"\)" apps/api/routes -S || trueRepository: Iro96/Mind-Bus
Length of output: 11289
🏁 Script executed:
#!/bin/bash
set -euo pipefail
# Locate delete_memory implementation
rg -n --type=py "def delete_memory" -S memory apps/api || true
# Show the implementation block
fd -e py | rg -n --type=py "def delete_memory" -S . | head -n 20 || true
# Print the relevant section(s)
# Use rg with context to capture surrounding code
rg -n --type=py "def delete_memory\(" -C3 memory || true
rg -n --type=py "delete_memory\(" -C3 memory || trueRepository: Iro96/Mind-Bus
Length of output: 298
🏁 Script executed:
#!/bin/bash
set -euo pipefail
# Find LongTermMemoryManager class definition
rg -n --type=py "class LongTermMemoryManager" -S memory apps/api || true
# Show delete_memory method inside long_term.py
rg -n --type=py "def delete_memory" -S memory apps/api || true
# Print the long_term.py delete_memory method body with some context
rg -n --type=py "def delete_memory" -C3 memory/long_term.py || true
# If multiple, show full surrounding class region for any match
rg -n --type=py "def delete_memory" -C20 memory/long_term.py || trueRepository: Iro96/Mind-Bus
Length of output: 229
🏁 Script executed:
#!/bin/bash
set -euo pipefail
# Get line number for delete_memory in long_term.py
rg -n --type=py "def delete_memory" memory/long_term.py || true
# Show a wider region around the definition (or entire file if small)
show_lines() {
local file="$1"; local start="$2"; local end="$3"
awk -v s="$start" -v e="$end" 'NR>=s && NR<=e {printf "%6d\t%s\n", NR, $0}' "$file"
}
line=$(rg -n --type=py "def delete_memory" memory/long_term.py | head -n1 | cut -d: -f1 || true)
if [ -n "${line:-}" ]; then
start=$((line-30))
end=$((line+120))
[ $start -lt 1 ] && start=1
show_lines memory/long_term.py $start $end
else
echo "No delete_memory found in memory/long_term.py"
fiRepository: Iro96/Mind-Bus
Length of output: 104
🏁 Script executed:
#!/bin/bash
set -euo pipefail
rg -n --type=py "delete_memory" memory || true
# Show any matching function/method definitions
rg -n --type=py "def delete_memory" -S memory || true
# Also search for UUID(memory_id) or "Failed to delete" messages
rg -n --type=py "UUID\(memory_id\)|Failed to delete memory" -S . || trueRepository: Iro96/Mind-Bus
Length of output: 273
🏁 Script executed:
#!/bin/bash
set -euo pipefail
wc -l memory/long_term.py
echo "---- memory/long_term.py ----"
sed -n '1,240p' memory/long_term.pyRepository: Iro96/Mind-Bus
Length of output: 7707
🏁 Script executed:
#!/bin/bash
set -euo pipefail
# Check if any other route decorators use empty string or root path
rg -n --type=py "`@router`\.(get|post|put|patch|delete)\(\"(\"\s*,|\")" apps/api/routes || true
rg -n --type=py "`@router`\.(get|post|put|patch|delete)\(\"/\"\s*,|\")" apps/api/routes || true
# See if any router instances are created with a prefix
rg -n --type=py "APIRouter\(\s*prefix\s*=" apps/api/routes || true
# List all include_router(...) calls that might include prefixes
rg -n --type=py "include_router\(" apps/api/main.py apps/api/routes -S || true
# Search for other single-segment DELETE routes outside memory
rg -n --type=py "`@router`\.delete\(\"\ /\{[^/}]+\}\"\)" apps/api/routes || true
rg -n --type=py "`@router`\.delete\(\"\(/\{[^/}]+\}\"\)" apps/api/routes || true
rg -n --type=py "`@router`\.delete\(\"/\{[^/}]+\}\"\)" apps/api/routes || trueRepository: Iro96/Mind-Bus
Length of output: 1128
🏁 Script executed:
#!/bin/bash
set -euo pipefail
# Show all APIRouter(prefix=...) declarations
echo "== Router prefixes in apps/api/routes/*.py =="
rg -n --type=py "router = APIRouter\(.*prefix\s*=" apps/api/routes -S || true
# List routes with path pattern "/{...}" (single-segment dynamic) across apps/api/routes
echo
echo "== Root-style dynamic routes '`@router`.(get|post|... )('/{...}')' =="
rg -n --type=py "`@router`\.(get|post|put|patch|delete)\(/\"\{\w+[^}]*\}\"\)" apps/api/routes -S || true
rg -n --type=py "`@router`\.(get|post|put|patch|delete)\(\"/\{[^/}]+\}\"\)" apps/api/routes -S || true
# Specifically find any routers with no prefix (router = APIRouter() without prefix)
echo
echo "== Routers declared without prefix (router = APIRouter()) =="
rg -n --type=py "router\s*=\s*APIRouter\(\s*\)" apps/api/routes -S || true
# Check tools.py route base for `@router.get`("") to understand join behavior pattern
echo
echo "== apps/api/routes/tools.py `@router.get`(\"\") context =="
rg -n --type=py "`@router`\.get\(\"\"\)" -C3 apps/api/routes/tools.py || trueRepository: Iro96/Mind-Bus
Length of output: 1028
🏁 Script executed:
#!/bin/bash
set -euo pipefail
# Find any route decorators that define a single-segment dynamic path at the router level.
# (i.e., "/{something}") inside whatever router that file declares.
rg -n --type=py '`@router`\.(get|post|put|patch|delete)\("/\{[^/}]+\}"\)' apps/api/routes -S || true
# Also catch variations like "/{something}" with spaces
rg -n --type=py '`@router`\.(get|post|put|patch|delete)\("/\{\s*[^/}]+\s*\}"\)' apps/api/routes -S || true
# Show the first few lines of documents.py and chat.py route declarations to see their paths
echo
echo "== documents.py top routes =="
rg -n --type=py '`@router`\.(get|post|put|patch|delete)\("' apps/api/routes/documents.py || true
echo
echo "== chat.py top routes =="
rg -n --type=py '`@router`\.(get|post|put|patch|delete)\("' apps/api/routes/chat.py || trueRepository: Iro96/Mind-Bus
Length of output: 709
🏁 Script executed:
#!/bin/bash
set -euo pipefail
# Find calls to memories endpoints in repo (clients/tests/etc.)
rg -n --type=py "('/memories|\"/memories|/memory|\"/memory|DELETE.*memories|delete.*memories|memories/refresh|/documents" . || true
# Specifically look for delete memory usage
rg -n --type=py "delete.*memory|/memory/|/memories/|memories/{|memories/" tests apps . || true
# Confirm there is no other memory delete route
rg -n --type=py "`@router`\.delete\(" apps/api/routes | rg -n "memory|memories" || trueRepository: Iro96/Mind-Bus
Length of output: 475
🏁 Script executed:
#!/bin/bash
set -euo pipefail
echo "== Search for literal '/memories' usage =="
rg -n --type=py "/memories" . || true
rg -n --type=js "/memories" . || true
rg -n --type=ts "/memories" . || true
rg -n --type=tsx "/memories" . || true
rg -n --type=jsx "/memories" . || true
rg -n --type=md "/memories" . || true
echo
echo "== Search for literal '/memory' usage =="
rg -n --type=py "/memory" . || true
rg -n --type=js "/memory" . || true
rg -n --type=ts "/memory" . || true
rg -n --type=tsx "/memory" . || true
rg -n --type=md "/memory" . || true
echo
echo "== Search for 'memories/refresh' usage =="
rg -n --type=py "memories/refresh" . || true
rg -n --type=js "memories/refresh" . || true
rg -n --type=ts "memories/refresh" . || true
rg -n --type=md "memories/refresh" . || trueRepository: Iro96/Mind-Bus
Length of output: 1309
🏁 Script executed:
#!/bin/bash
set -euo pipefail
# 1) Confirm frontend calls + expected paths
echo "== Frontend calls around MemoryView.vue =="
nl -ba frontend/src/views/MemoryView.vue 70 140 2>/dev/null || sed -n '70,140p' frontend/src/views/MemoryView.vue
echo
echo "== Frontend route config (router/index.ts) =="
sed -n '1,80p' frontend/src/router/index.ts
# 2) Confirm docs expect /memory endpoints
echo
echo "== docs/INTEGRATION.md memory section =="
rg -n "/memory|DELETE /memory" docs/INTEGRATION.md
sed -n '160,230p' docs/INTEGRATION.md
# 3) Search for any delete_memory implementation anywhere
echo
echo "== Search for delete_memory definitions anywhere in repo =="
rg -n --type=py "def delete_memory" . || true
# 4) Search for 'delete_memory(' calls besides route
echo
echo "== Search for delete_memory( calls =="
rg -n --type=py "delete_memory\s*\(" . || trueRepository: Iro96/Mind-Bus
Length of output: 4923
Fix memory router mounting: /memory alias and delete route resolve to the app root
apps/api/main.pymountsmemory.routerwith no prefix (app.include_router(memory.router)), andapps/api/routes/memory.pyusesrouter = APIRouter()with@router.get("")/@router.delete("/{memory_id}"), so these becomeGET /andDELETE /{memory_id}—notGET /memoryandDELETE /memory/{memory_id}as expected by the frontend (frontend/src/views/MemoryView.vue) anddocs/INTEGRATION.md.delete_memorycallsmemory_manager.delete_memory(...), butLongTermMemoryManager(memory/long_term.py) has nodelete_memorymethod, so the handler will fail (HTTP 500) even after fixing the route.
🧰 Tools
🪛 Ruff (0.15.14)
[warning] 169-169: Do not perform function call Depends in argument defaults; instead, perform the call within the function, or read the default from a module-level singleton variable
(B008)
[warning] 192-192: Do not perform function call Depends in argument defaults; instead, perform the call within the function, or read the default from a module-level singleton variable
(B008)
🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@apps/api/routes/memory.py` around lines 165 - 192, The router is mounted
without a prefix so `@router.get`("") and `@router.delete`("/{memory_id}") become
GET / and DELETE /{memory_id}; add a "/memory" prefix (either change APIRouter()
to APIRouter(prefix="/memory") in the module that defines router or
include_router(memory.router, prefix="/memory") where the router is mounted) so
get_memories_alias remains GET /memory and delete_memory becomes DELETE
/memory/{memory_id}; also implement the missing delete_memory method on
LongTermMemoryManager in memory/long_term.py (matching the signature used by the
handler) or change the handler to call the existing removal method name—ensure
the handler calls a real LongTermMemoryManager API (referenced symbols:
APIRouter, get_memories_alias, delete_memory handler,
memory_manager.delete_memory, LongTermMemoryManager).
Comment on lines
+117
to
+119
| onUploadProgress: (e) => { | ||
| uploadProgress.value = Math.round((e.loaded / e.total) * 100) | ||
| } |
There was a problem hiding this comment.
🧩 Analysis chain
🌐 Web query:
Is the total property of axios AxiosProgressEvent optional in axios 1.5?
💡 Result:
Yes, the total property of AxiosProgressEvent is optional in axios 1.5 [1]. In the TypeScript type definition for AxiosProgressEvent, the total property is marked as optional (denoted by the? symbol), meaning it can be undefined [1]. This is because the total size of data (often derived from the Content-Length header) is not always available or computable during a transfer, such as in certain cases involving gzipped responses or streaming data where the length is unknown [2][3]. Official axios documentation and type definitions confirm this structure, identifying that developers should handle cases where total might be absent [1]. When using this property, it is recommended to verify that it exists or that lengthComputable is true before performing calculations to avoid errors or inaccurate progress reporting [4][3].
Citations:
- 1: https://github.com/axios/axios/blob/v1.15.0/docs/pages/advanced/progress-capturing.md
- 2: onDownloadProgress has progressEvent.total property with zero value axios/axios#1591
- 3: https://developer.mozilla.org/en-US/docs/Web/API/ProgressEvent/total
- 4: https://stackoverflow.com/questions/44936028/progress-bar-with-axios
🏁 Script executed:
#!/bin/bash
set -euo pipefail
FILE="frontend/src/views/DocumentsView.vue"
# Show the relevant region around the reported progress handler
nl -ba "$FILE" | sed -n '90,160p'
# Find any object URL creation and whether revokeObjectURL exists
echo "---- URL.createObjectURL occurrences ----"
rg -n "URL\.createObjectURL" "$FILE" || true
echo "---- URL.revokeObjectURL occurrences ----"
rg -n "URL\.revokeObjectURL" "$FILE" || true
# Find download filename handling
echo "---- link.download occurrences ----"
rg -n "download\s*=" "$FILE" || true
# Find isDragging usage
echo "---- isDragging occurrences ----"
rg -n "isDragging" "$FILE" || trueRepository: Iro96/Mind-Bus
Length of output: 101
🏁 Script executed:
#!/bin/bash
set -euo pipefail
FILE="frontend/src/views/DocumentsView.vue"
echo "---- Progress region (approx 90-160) ----"
cat -n "$FILE" | sed -n '90,160p'
echo "---- URL.createObjectURL occurrences ----"
rg -n "URL\.createObjectURL" "$FILE" || true
echo "---- URL.revokeObjectURL occurrences ----"
rg -n "URL\.revokeObjectURL" "$FILE" || true
echo "---- link.download occurrences ----"
rg -n "link\.download|download\s*=" "$FILE" || true
echo "---- isDragging occurrences ----"
rg -n "isDragging" "$FILE" || trueRepository: Iro96/Mind-Bus
Length of output: 2793
🏁 Script executed:
#!/bin/bash
set -euo pipefail
FILE="frontend/src/views/DocumentsView.vue"
cat -n "$FILE" | sed -n '1,120p'Repository: Iro96/Mind-Bus
Length of output: 4468
🏁 Script executed:
#!/bin/bash
set -euo pipefail
FILE="frontend/src/views/DocumentsView.vue"
cat -n "$FILE" | sed -n '108,160p'Repository: Iro96/Mind-Bus
Length of output: 1835
Guard upload progress against undefined e.total (and fix download URL cleanup/filename)
The current onUploadProgress computes e.loaded / e.total; in axios AxiosProgressEvent.total is optional, so e.total can be undefined and produce NaN%.
onUploadProgress: (e) => {
uploadProgress.value = Math.round((e.loaded / e.total) * 100)
}- Revoke the created blob URL after download (
window.URL.revokeObjectURL(url)) to avoid leaking object URLs. - Preserve the filename:
link.download = ''forces a generic name—set it fromdoc.filename(or useContent-Dispositionif the backend provides it). isDraggingis always forced tofalse(never set totrue), so it’s dead state unless there’s styling/UX that’s missing.
🐛 Proposed fix
onUploadProgress: (e) => {
- uploadProgress.value = Math.round((e.loaded / e.total) * 100)
+ if (e.total != null) {
+ uploadProgress.value = Math.round((e.loaded / e.total) * 100)
+ }
}🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
In `@frontend/src/views/DocumentsView.vue` around lines 117 - 119, Guard the
upload progress handler in onUploadProgress so it handles a possibly undefined
e.total (compute percent only when e.total is a number, default to 0 or e.loaded
when unknown) and assign the result to uploadProgress.value; after creating a
blob URL for download, call window.URL.revokeObjectURL(url) once the link is
clicked or after triggering the download to avoid leaking object URLs, and set
link.download to doc.filename (or use a parsed filename from
Content-Disposition) instead of an empty string so the original filename is
preserved; also review the isDragging state (it’s never set true) and either
remove it or wire it to the drag events to make it meaningful.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Release Notes
New Features
Documentation
Chores