feat: Bug Recommendation system using vector cosine-similarity#1
Open
SuhaniBhati wants to merge 3 commits intomainfrom
Open
feat: Bug Recommendation system using vector cosine-similarity#1SuhaniBhati wants to merge 3 commits intomainfrom
SuhaniBhati wants to merge 3 commits intomainfrom
Conversation
📝 WalkthroughWalkthroughAdds a Python Flask microservice that returns text embeddings and updates the Node.js core service with vector creation, an external embedding integration, cosine similarity utilities, model schema changes, and recommendation endpoints for similar bugs. Changes
Sequence Diagram(s)sequenceDiagram
participant Client
participant CoreAPI as Core API (Node.js)
participant AIService as AI Service (Flask)
participant DB as Database (MongoDB)
Client->>CoreAPI: POST /api/vectors (bugId)
CoreAPI->>DB: Bug.findById(bugId)
DB-->>CoreAPI: bug document
CoreAPI->>AIService: POST /embed (built text)
AIService-->>CoreAPI: {embedding: [...]}
CoreAPI->>DB: Vector.findOneAndUpdate(... upsert)
DB-->>CoreAPI: saved vector
CoreAPI-->>Client: 200 {data: vector}
sequenceDiagram
participant Client
participant CoreAPI as Core API (Node.js)
participant DB as Database (MongoDB)
Client->>CoreAPI: GET /api/recommendations/:id
CoreAPI->>DB: Vector.findById(id)
DB-->>CoreAPI: target vector
CoreAPI->>DB: Vector.find(...) (other vectors)
DB-->>CoreAPI: other vectors
Note over CoreAPI: compute cosineSimilarity for each vector\nsort desc, pick top-5
CoreAPI->>DB: Bug.find({_id: {$in: topIds}})
DB-->>CoreAPI: bug documents
CoreAPI-->>Client: 200 {recommendations: [{bug, similarityScore}, ...]}
Estimated code review effort🎯 4 (Complex) | ⏱️ ~60 minutes Poem
🚥 Pre-merge checks | ✅ 2 | ❌ 1❌ Failed checks (1 warning)
✅ Passed checks (2 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 |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 6
🧹 Nitpick comments (5)
ai-service/requirements.txt (1)
1-4: Use pinned or bounded versions for runtime dependencies.Current bare package names make deployments non-reproducible across environments.
♻️ Suggested change
-flask -sentence-transformers -numpy -scikit-learn +flask>=3.0,<4.0 +sentence-transformers>=3.0,<4.0 +numpy>=2.0,<3.0 +scikit-learn>=1.5,<2.0🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@ai-service/requirements.txt` around lines 1 - 4, The requirements.txt currently lists unpinned runtime dependencies ("flask", "sentence-transformers", "numpy", "scikit-learn"); update each entry to use either exact pinned versions (e.g., "package==x.y.z") or conservative version bounds (e.g., "package>=x.y,<x+1") to make deployments reproducible. Determine compatible versions used in CI or local environment and replace the bare names in requirements.txt with those pinned/bounded specifications so installs are deterministic.core-services/src/models/Vector.js (1)
4-25: Explicit_id: ObjectIdfield declaration is unnecessary.Mongoose auto-generates
_idas an ObjectId by default. Declaring it explicitly withoutauto: falsehas no effect and may cause confusion. Consider removing line 5.🧹 Proposed cleanup
const vectorSchema = new mongoose.Schema({ - _id: ObjectId, bugId: {🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@core-services/src/models/Vector.js` around lines 4 - 25, Remove the explicit _id declaration from the mongoose schema to avoid confusion since Mongoose auto-generates an ObjectId; update the vectorSchema by deleting the "_id: ObjectId" entry so the schema only contains bugId, vector, modelVersion, and updatedAt, leaving default Mongoose _id behavior intact (refer to the vectorSchema constant to locate the schema definition).core-services/src/services/embedding.Service.js (1)
5-18: Consider adding a request timeout.The axios request has no timeout configured. If the AI service is unresponsive, this call will hang indefinitely, potentially exhausting connection pools or blocking the request thread.
⏱️ Proposed fix: add timeout
const res = await axios.post( `${process.env.AI_SERVICE_URL}/embed`, { title: bug.title, description: bug.description, - tags: Array.isArray(bug.tags) - ? bug.tags.join(", ") - : bug.tags, - techStack: Array.isArray(bug.techStack) - ? bug.techStack.join(", ") - : bug.techStack, + tags: Array.isArray(bug.tags) ? bug.tags : [], + techStack: Array.isArray(bug.techStack) ? bug.techStack : [], difficulty: bug.difficulty - } + }, + { timeout: 10000 } // 10 second timeout );🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@core-services/src/services/embedding.Service.js` around lines 5 - 18, The axios.post call in embedding.Service.js that sends payload from bug (title, description, tags, techStack, difficulty) lacks a timeout and can hang; add an axios request timeout (e.g., pass a config object with a timeout value in milliseconds as the third argument to axios.post) to the call that assigns to res so the request fails fast when the AI service is unresponsive, and choose a sensible timeout constant (e.g., DEFAULT_EMBED_TIMEOUT) used by the module or service.core-services/src/controllers/recommendation.Controller.js (1)
15-22: Full collection scan doesn't scale.
Vector.find({ bugId: { $ne: bugId } })loads all vectors into memory and MongoDB cannot efficiently use indexes for$nequeries. This is acceptable for a small dataset but will degrade as the collection grows.For future scalability, consider:
- Using MongoDB Atlas Vector Search or a dedicated vector database
- Implementing pagination or limiting candidates
- Pre-computing nearest neighbors offline
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@core-services/src/controllers/recommendation.Controller.js` around lines 15 - 22, The current Vector.find({ bugId: { $ne: bugId } }) causes a full collection scan and will not scale; change the recommendation logic in the recommendation controller to avoid loading all vectors into memory by limiting candidate vectors (e.g., add a .limit(N) and sensible sort/filters or sample recent / relevant documents) or delegate to a proper vector search (MongoDB Atlas Vector Search or an external vector DB) instead of mapping all results; update the code around the Vector.find call and the downstream similarities/top5 computation so it queries a bounded candidate set (use Vector.find(query).sort(...).limit(CANDIDATE_LIMIT) or replace with a vector-search API call) and ensure cosineSimilarity and top5 logic operate only on that limited result set.core-services/src/controllers/vector.Controller.js (1)
24-35: Consider validating the embedding before storing.If the AI service returns an unexpected response (null, empty array, wrong dimensions), this will store an invalid vector. A quick validation could prevent corrupted data.
🛡️ Proposed validation
const embedding = await getEmbedding(bug); + if (!Array.isArray(embedding) || embedding.length === 0) { + return res.status(502).json({ message: "Invalid embedding from AI service" }); + } + const vectorDoc = await Vector.findOneAndUpdate(🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@core-services/src/controllers/vector.Controller.js` around lines 24 - 35, The code calls getEmbedding(bug) then upserts it via Vector.findOneAndUpdate without validation; add a validation step after getEmbedding to ensure the returned embedding is non-null, is an Array of numbers, and matches the expected dimensionality for "all-MiniLM-L6-v2" (e.g., expectedLength) before calling Vector.findOneAndUpdate; if validation fails, log or throw a clear error (including bug._id and the embedding shape) and skip the DB write so invalid vectors are not stored. Use the symbols getEmbedding, Vector.findOneAndUpdate, and modelVersion in your checks and error messages.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@ai-service/app.py`:
- Around line 9-12: Check that request.json is present and valid before calling
build_text; if request.json is None or missing required fields, return a 400 Bad
Request instead of proceeding. Modify the handler around request.json /
build_text(data) to validate data (e.g., ensure data is not None and has
expected keys) and return an appropriate error response via jsonify with a 400
status; only call build_text(data) and get_embedding(text) when validation
passes.
- Line 15: The app currently calls app.run(port=3000, debug=True) which
hardcodes Flask debug mode; remove the hardcoded debug=True and make debug
configurable via environment/config (e.g., read from an env var or app config)
so debug is only enabled in development; update the app.run call in app.py to
use that config value (e.g., app.config or os.environ) and ensure the default is
False in production.
In `@ai-service/utils.py`:
- Around line 5-9: Normalize the incoming tags and techStack values before
joining: ensure tags and tech_stack are lists (convert None to [], wrap a string
into a single-item list rather than iterating its characters, and leave existing
lists untouched) before constructing tags_text and tech_text. Update the
variables tags and tech_stack in ai-service/utils.py so they handle str, None,
and list types safely (e.g., check isinstance(tags, str) /
isinstance(tech_stack, str) and convert accordingly) and then use " ".join on
the normalized lists to produce tags_text and tech_text.
In `@core-services/src/controllers/recommendation.Controller.js`:
- Line 7: The controller reads const bugId = req.params.id without validating
it; add an ObjectId format check (e.g., using mongoose.Types.ObjectId.isValid)
at the start of the relevant handler in recommendation.Controller.js (where
bugId is used) and return a 400 Bad Request with a clear message when invalid,
so invalid ids do not trigger Mongoose cast errors; keep the existing variable
name bugId and perform the check before any DB calls that use bugId.
In `@core-services/src/services/embedding.Service.js`:
- Around line 10-15: The payload currently converts arrays to comma-separated
strings for the tags and techStack fields, which breaks the Python build_text()
logic that expects arrays; update the assignments in embedding.Service.js so
tags and techStack are sent as arrays (e.g., for tags: use bug.tags directly if
Array.isArray(bug.tags) otherwise wrap/split into an array) and do the same for
techStack, ensuring the service sends arrays to match ai-service/utils.py's
build_text().
In `@core-services/src/utils/similarityScore.js`:
- Around line 1-13: In cosineSimilarity, validate inputs and guard against NaNs:
check that A and B are arrays and have the same length (or explicitly iterate to
Math.min(A.length, B.length)), coerce entries to numbers (skip/ treat undefined
as 0), compute norms and if either norm is zero return 0 (or a defined fallback)
instead of performing the division, and ensure you never divide by
Math.sqrt(normA)*Math.sqrt(normB); update exports.cosineSimilarity accordingly
to return a deterministic numeric value when vectors are zero-length,
zero-magnitude, or length-mismatched.
---
Nitpick comments:
In `@ai-service/requirements.txt`:
- Around line 1-4: The requirements.txt currently lists unpinned runtime
dependencies ("flask", "sentence-transformers", "numpy", "scikit-learn"); update
each entry to use either exact pinned versions (e.g., "package==x.y.z") or
conservative version bounds (e.g., "package>=x.y,<x+1") to make deployments
reproducible. Determine compatible versions used in CI or local environment and
replace the bare names in requirements.txt with those pinned/bounded
specifications so installs are deterministic.
In `@core-services/src/controllers/recommendation.Controller.js`:
- Around line 15-22: The current Vector.find({ bugId: { $ne: bugId } }) causes a
full collection scan and will not scale; change the recommendation logic in the
recommendation controller to avoid loading all vectors into memory by limiting
candidate vectors (e.g., add a .limit(N) and sensible sort/filters or sample
recent / relevant documents) or delegate to a proper vector search (MongoDB
Atlas Vector Search or an external vector DB) instead of mapping all results;
update the code around the Vector.find call and the downstream similarities/top5
computation so it queries a bounded candidate set (use
Vector.find(query).sort(...).limit(CANDIDATE_LIMIT) or replace with a
vector-search API call) and ensure cosineSimilarity and top5 logic operate only
on that limited result set.
In `@core-services/src/controllers/vector.Controller.js`:
- Around line 24-35: The code calls getEmbedding(bug) then upserts it via
Vector.findOneAndUpdate without validation; add a validation step after
getEmbedding to ensure the returned embedding is non-null, is an Array of
numbers, and matches the expected dimensionality for "all-MiniLM-L6-v2" (e.g.,
expectedLength) before calling Vector.findOneAndUpdate; if validation fails, log
or throw a clear error (including bug._id and the embedding shape) and skip the
DB write so invalid vectors are not stored. Use the symbols getEmbedding,
Vector.findOneAndUpdate, and modelVersion in your checks and error messages.
In `@core-services/src/models/Vector.js`:
- Around line 4-25: Remove the explicit _id declaration from the mongoose schema
to avoid confusion since Mongoose auto-generates an ObjectId; update the
vectorSchema by deleting the "_id: ObjectId" entry so the schema only contains
bugId, vector, modelVersion, and updatedAt, leaving default Mongoose _id
behavior intact (refer to the vectorSchema constant to locate the schema
definition).
In `@core-services/src/services/embedding.Service.js`:
- Around line 5-18: The axios.post call in embedding.Service.js that sends
payload from bug (title, description, tags, techStack, difficulty) lacks a
timeout and can hang; add an axios request timeout (e.g., pass a config object
with a timeout value in milliseconds as the third argument to axios.post) to the
call that assigns to res so the request fails fast when the AI service is
unresponsive, and choose a sensible timeout constant (e.g.,
DEFAULT_EMBED_TIMEOUT) used by the module or service.
🪄 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: 649e3593-65d8-42f9-9903-02378c34ea97
⛔ Files ignored due to path filters (1)
core-services/package-lock.jsonis excluded by!**/package-lock.json
📒 Files selected for processing (15)
.gitignoreai-service/app.pyai-service/model.pyai-service/requirements.txtai-service/utils.pycore-services/package.jsoncore-services/src/app.jscore-services/src/controllers/recommendation.Controller.jscore-services/src/controllers/vector.Controller.jscore-services/src/models/Bug.jscore-services/src/models/Vector.jscore-services/src/routes/recommendation.Routes.jscore-services/src/routes/vector.Routes.jscore-services/src/services/embedding.Service.jscore-services/src/utils/similarityScore.js
Comment on lines
+9
to
+12
| data = request.json | ||
| text = build_text(data) | ||
| embedding = get_embedding(text) | ||
| return jsonify({"embedding": embedding}) |
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
find . -name "app.py" -path "*/ai-service/*" | head -20Repository: call-0f-code/Bug-Graveyard
Length of output: 89
🏁 Script executed:
cat -n ai-service/app.pyRepository: call-0f-code/Bug-Graveyard
Length of output: 549
🏁 Script executed:
# Search for build_text function definition
rg "def build_text" -A 5Repository: call-0f-code/Bug-Graveyard
Length of output: 388
🏁 Script executed:
# Check Flask version if available
if [ -f "requirements.txt" ]; then cat requirements.txt; elif [ -f "setup.py" ]; then grep -i flask setup.py; elif [ -f "pyproject.toml" ]; then grep -i flask pyproject.toml; fiRepository: call-0f-code/Bug-Graveyard
Length of output: 52
Validate request body before building text.
If request JSON is missing or invalid, request.json returns None and the call to build_text(data) will raise an AttributeError when attempting to call .get() on None, resulting in a 500 error instead of a proper 400 Bad Request.
🐛 Proposed fix
`@app.route`("/embed", methods=["POST"])
def embed():
- data = request.json
+ data = request.get_json(silent=True)
+ if not isinstance(data, dict):
+ return jsonify({"error": "Invalid JSON payload"}), 400
text = build_text(data)
embedding = get_embedding(text)
return jsonify({"embedding": embedding})🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@ai-service/app.py` around lines 9 - 12, Check that request.json is present
and valid before calling build_text; if request.json is None or missing required
fields, return a 400 Bad Request instead of proceeding. Modify the handler
around request.json / build_text(data) to validate data (e.g., ensure data is
not None and has expected keys) and return an appropriate error response via
jsonify with a 400 status; only call build_text(data) and get_embedding(text)
when validation passes.
Comment on lines
+5
to
+9
| tags = data.get("tags", []) | ||
| tech_stack = data.get("techStack", []) | ||
| difficulty = data.get("difficulty", "") | ||
| tags_text = " ".join(tags) | ||
| tech_text = " ".join(tech_stack) |
There was a problem hiding this comment.
Normalize tags/techStack before joining.
Line 8 and Line 9 break when values are strings or null: strings become character-spaced text, and None raises TypeError. This is likely with the current upstream payload shape from core-services/src/services/embedding.Service.js:10-15.
🐛 Proposed fix
def build_text(data):
# Safe extraction
title = data.get("title", "")
description = data.get("description", "")
tags = data.get("tags", [])
tech_stack = data.get("techStack", [])
difficulty = data.get("difficulty", "")
- tags_text = " ".join(tags)
- tech_text = " ".join(tech_stack)
+ def normalize_terms(value):
+ if value is None:
+ return ""
+ if isinstance(value, (list, tuple, set)):
+ return " ".join(str(v).strip() for v in value if v is not None)
+ return str(value).replace(",", " ").strip()
+
+ tags_text = normalize_terms(tags)
+ tech_text = normalize_terms(tech_stack)📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
Suggested change
| tags = data.get("tags", []) | |
| tech_stack = data.get("techStack", []) | |
| difficulty = data.get("difficulty", "") | |
| tags_text = " ".join(tags) | |
| tech_text = " ".join(tech_stack) | |
| tags = data.get("tags", []) | |
| tech_stack = data.get("techStack", []) | |
| difficulty = data.get("difficulty", "") | |
| def normalize_terms(value): | |
| if value is None: | |
| return "" | |
| if isinstance(value, (list, tuple, set)): | |
| return " ".join(str(v).strip() for v in value if v is not None) | |
| return str(value).replace(",", " ").strip() | |
| tags_text = normalize_terms(tags) | |
| tech_text = normalize_terms(tech_stack) |
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@ai-service/utils.py` around lines 5 - 9, Normalize the incoming tags and
techStack values before joining: ensure tags and tech_stack are lists (convert
None to [], wrap a string into a single-item list rather than iterating its
characters, and leave existing lists untouched) before constructing tags_text
and tech_text. Update the variables tags and tech_stack in ai-service/utils.py
so they handle str, None, and list types safely (e.g., check isinstance(tags,
str) / isinstance(tech_stack, str) and convert accordingly) and then use "
".join on the normalized lists to produce tags_text and tech_text.
…atch , improve cosine similarity
There was a problem hiding this comment.
Actionable comments posted: 2
🧹 Nitpick comments (1)
ai-service/app.py (1)
20-22: Consider using theloggingmodule instead ofprint().For production services, using Python's
loggingmodule provides better control over log levels, formatting, and output destinations.♻️ Suggested improvement
+import logging import os from dotenv import load_dotenv ... +logging.basicConfig(level=logging.INFO) +logger = logging.getLogger(__name__) app = Flask(__name__) ... except Exception as e: - print("Error in /embed:", str(e)) + logger.exception("Error in /embed") return jsonify({"error": "Internal server error"}), 500🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@ai-service/app.py` around lines 20 - 22, Replace the ad-hoc print in the /embed exception handler with structured logging: import and configure Python's logging (or use Flask's app.logger) and call logger.exception("Error in /embed") (or app.logger.exception) to record the error and stack trace instead of print(str(e)); keep the existing jsonify return and status code. Reference the exception handler block for /embed and the exception variable e when making the change.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@ai-service/app.py`:
- Around line 24-27: The DEBUG assignment currently compares a string to a list
(DEBUG = os.getenv("FLASK_DEBUG", "false").lower() == ["true", "1", "yes"])
which is always False; change it to check membership instead by evaluating the
lowered env value against the list of truthy strings (use "in" with the list) so
DEBUG correctly reflects FLASK_DEBUG; update the DEBUG expression near PORT and
app.run references (symbols: PORT, DEBUG, app.run, os.getenv("FLASK_DEBUG")) to
use membership testing.
- Around line 16-19: The duplicate call to build_text(data) is unnecessary and
wastes CPU; remove the second invocation so the handler only calls
build_text(data) once to produce text, then passes that text to
get_embedding(text) and returns jsonify({"embedding": embedding}), 200; locate
the duplicate call in the same block that calls build_text and get_embedding and
delete the redundant build_text(data) invocation.
---
Nitpick comments:
In `@ai-service/app.py`:
- Around line 20-22: Replace the ad-hoc print in the /embed exception handler
with structured logging: import and configure Python's logging (or use Flask's
app.logger) and call logger.exception("Error in /embed") (or
app.logger.exception) to record the error and stack trace instead of
print(str(e)); keep the existing jsonify return and status code. Reference the
exception handler block for /embed and the exception variable e when making the
change.
🪄 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: fbdb4142-eeb7-47f5-95c3-5850846874a8
📒 Files selected for processing (6)
ai-service/app.pyai-service/requirements.txtai-service/utils.pycore-services/src/controllers/recommendation.Controller.jscore-services/src/services/embedding.Service.jscore-services/src/utils/similarityScore.js
✅ Files skipped from review due to trivial changes (1)
- ai-service/requirements.txt
🚧 Files skipped from review as they are similar to previous changes (4)
- core-services/src/utils/similarityScore.js
- core-services/src/services/embedding.Service.js
- ai-service/utils.py
- core-services/src/controllers/recommendation.Controller.js
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (1)
ai-service/app.py (1)
19-21: Replaceprint()with proper logging.Using
print()for error output is not production-ready. Theloggingmodule provides better control over log levels, formatting, and output destinations, making it easier to monitor and debug the service in production environments.📋 Proposed refactor to use logging module
Add at the top of the file after imports:
+import logging + +logging.basicConfig(level=logging.INFO) +logger = logging.getLogger(__name__)Then update the exception handler:
except Exception as e: - print("Error in /embed:", str(e)) + logger.exception("Error in /embed endpoint") return jsonify({"error": "Internal server error"}), 500🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@ai-service/app.py` around lines 19 - 21, Replace the use of print() in the /embed exception handler with the logging module: import and configure a logger (or use Flask's app.logger) at top of ai-service/app.py, and inside the except Exception as e block replace print("Error in /embed:", str(e)) with a logger.error call that includes context ("Error in /embed") and the exception (e) so the stacktrace and message are recorded at error level; ensure logging is initialized before the route is registered.
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@ai-service/app.py`:
- Around line 23-26: Validate parsing of the PORT environment variable in the
__main__ block: replace the direct int(os.getenv("PORT", 3000)) usage with code
that attempts to parse os.getenv("PORT") inside a try/except (or use str.isdigit
checks), on ValueError log a clear error (use the app or module logger), fall
back to the default port 3000 (or exit if you prefer strict behavior), and then
call app.run(port=PORT, debug=DEBUG); update the section around the PORT
variable assignment and the app.run call to use the validated PORT.
---
Nitpick comments:
In `@ai-service/app.py`:
- Around line 19-21: Replace the use of print() in the /embed exception handler
with the logging module: import and configure a logger (or use Flask's
app.logger) at top of ai-service/app.py, and inside the except Exception as e
block replace print("Error in /embed:", str(e)) with a logger.error call that
includes context ("Error in /embed") and the exception (e) so the stacktrace and
message are recorded at error level; ensure logging is initialized before the
route is registered.
🪄 Autofix (Beta)
Fix all unresolved CodeRabbit comments on this PR:
- Push a commit to this branch (recommended)
- Create a new PR with the fixes
Comment on lines
+23
to
+26
| if __name__ == "__main__": | ||
| PORT = int(os.getenv("PORT", 3000)) | ||
| DEBUG = os.getenv("FLASK_DEBUG", "false").lower() in ["true", "1", "yes"] | ||
| app.run(port=PORT, debug=DEBUG) No newline at end of file |
There was a problem hiding this comment.
Add error handling for PORT environment variable parsing.
If the PORT environment variable contains a non-numeric value, int() will raise a ValueError and crash the application at startup. While this is an edge case, adding validation would make the service more robust.
🛡️ Proposed fix to handle invalid PORT values
if __name__ == "__main__":
- PORT = int(os.getenv("PORT", 3000))
+ try:
+ PORT = int(os.getenv("PORT", "3000"))
+ except ValueError:
+ PORT = 3000
DEBUG = os.getenv("FLASK_DEBUG", "false").lower() in ["true", "1", "yes"]
app.run(port=PORT, debug=DEBUG)🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@ai-service/app.py` around lines 23 - 26, Validate parsing of the PORT
environment variable in the __main__ block: replace the direct
int(os.getenv("PORT", 3000)) usage with code that attempts to parse
os.getenv("PORT") inside a try/except (or use str.isdigit checks), on ValueError
log a clear error (use the app or module logger), fall back to the default port
3000 (or exit if you prefer strict behavior), and then call app.run(port=PORT,
debug=DEBUG); update the section around the PORT variable assignment and the
app.run call to use the validated PORT.
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.
Summary by CodeRabbit