- Execute SELECT 1 against PostgreSQL via SQLAlchemy session to verify DB connectivity
- HTTP GET to ${MLFLOW_TRACKING_URI}/health (default: http://localhost:5000/health) to verify MLflow
- Return HTTP 503 if any component is unhealthy, HTTP 200 only when both are healthy
- Values: "healthy" / "unhealthy" for database and mlflow fields
- Add `import requests as http_requests` and `sa_text` imports
Closes task 5.6 in openspec/changes/code-review-fix/tasks.md
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Sort incoming candles by their time field in ascending chronological order
before processing to ensure correct feature engineering regardless of input
order. Also validate that no duplicate timestamps are present, returning
HTTP 400 with details if duplicates are found.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Parse start_date and end_date as datetime objects
- Return HTTP 400 if end_date is before start_date
- Return HTTP 400 if date range exceeds 365 days (1 year)
Closes task 5.4 in code-review-fix tasks.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
In /predict and /predict/batch endpoints, grab the model reference under
_model_swap_lock before running inference. Inference itself runs outside
the lock (using a local variable) to avoid blocking model swaps during
potentially slow computation.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Add verify_model_checksum() that validates model files against a
models/checksums.sha256 manifest before loading. Fails open when
manifest is missing or file not listed (backward compat), raises
HTTP 500 on hash mismatch. Created empty manifest placeholder.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Replace all instances of `detail=str(e)`, `detail=f"...{exc}"`, and similar
patterns that exposed internal exception messages to HTTP clients in
services/ml/app/main.py. All exception details are now logged server-side
only via logger.error(), while clients receive a generic "Internal server error"
message. Fixes 9 handlers across predict, batch predict, pattern detection,
training start, training runs fetch, training run delete, dataset info,
build dataset, and model load endpoints.
Mark task 5.1 as completed in tasks.md.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Added HTTP error checks before calling .json() in the three fetch
functions (fetchCandles, fetchAnnotations, fetchAnnotationTypes)
to prevent silent failures on non-2xx responses.
Marks task 4.12 as complete in tasks.md.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Guard all four fetch() calls in src/app/page.tsx against non-2xx HTTP
responses by throwing before attempting to parse the body as JSON.
Affected functions: fetchCharts, fetchAnnotations, fetchSpanAnnotations,
fetchSpanLabelTypes.
Marks task 4.11 as completed in code-review-fix/tasks.md.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Add sanitizeCsvCell() helper to both export routes that prefixes cell
values starting with =, +, @, or - with a single quote to prevent CSV
formula injection attacks.
Applied to:
- src/app/api/export/route.ts: timestamp and label_type columns
- src/app/api/span-annotations/export/route.ts: start_time, end_time,
label, and outcome columns
Closes task 4.10.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Import spanAnnotations from schema
- Wrap all delete operations in db.transaction() for atomicity
- Delete spanAnnotations first to satisfy FK constraints, then annotations, candles, chart
- Mark task 4.8 as done in tasks.md
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Reject DELETE ?all=true without chartId with HTTP 400 to prevent
accidental deletion of annotations across all charts.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Import z from zod
- Add CandleSchema validating time, open, high, low, close (number), volume (optional number)
- Add PatternDetectRequestSchema validating candles array and patterns array of non-empty strings
- Use safeParse() and return HTTP 400 with error details on validation failure
- Forward only validated data to the inference service
- Mark task 4.5 as completed in tasks.md
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Validates model_type as a non-empty string using .safeParse(); returns
HTTP 400 with error details on invalid input. Marks task 4.4 as done.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Validate run_id in POST /api/model/load using Zod:
- run_id must be a non-empty string matching /^[a-zA-Z0-9_-]+$/
- Returns HTTP 400 with error details if validation fails
- Validated data is forwarded to the inference service
Marks task 4.3 as complete in tasks.md.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Add BatchPredictRequestSchema with Zod to validate pair, timeframe,
start_date, and end_date fields. Returns HTTP 400 with flattened error
details on invalid input. Forward only validated data to the inference
service.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add CandleSchema validating time, open, high, low, close (number) and optional volume
- Add PredictRequestSchema validating pair (non-empty string), timeframe (non-empty string), candles array
- Use safeParse() and return HTTP 400 with error details on invalid input
- Forward only validated data to the inference service
- Mark task 4.1 as done in tasks.md
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add API_KEY environment variable with placeholder value 'change_me_to_a_strong_random_key'
- Include helpful comment explaining its purpose: authentication between Next.js and ML service
- Provide command for generating strong random value: openssl rand -hex 32
- Mark task 3.4 as completed
All 12 Next.js API routes that proxy requests to the ML service
(INFERENCE_API_URL / localhost:8001) now include the X-API-Key header
read from process.env.API_KEY. Affected routes:
- /api/predict
- /api/predict/batch
- /api/model/info
- /api/model/load
- /api/training/start
- /api/training/runs
- /api/training/runs/[run_id] (DELETE)
- /api/training/dataset-info
- /api/training/active
- /api/training/build-dataset
- /api/patterns/available
- /api/patterns/detect
Marks task 3.3 as complete in openspec/changes/code-review-fix/tasks.md.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Import Header, Depends, Security from fastapi
- Add verify_api_key dependency: reads API_KEY env var, checks X-API-Key
header, raises HTTP 401 if key mismatch; fail-open if env var not set
- Apply Depends(verify_api_key) to all 14 non-health endpoints
- /health endpoint remains unauthenticated for liveness probes
- Mark task 3.2 as complete in tasks.md
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Create src/middleware.ts with Next.js middleware
- Reads API_KEY env var and checks X-API-Key header on all /api/* routes
- Skips auth for /api/health endpoint
- Fails open (with warning) when API_KEY is not configured
- Returns 401 Unauthorized when key is missing or mismatched
- Mark task 3.1 as complete in tasks.md
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Replace hardcoded allow_origins=['*'] with dynamic configuration
- Read CORS_ORIGINS environment variable (comma-separated list)
- Default to 'http://localhost:3000' if CORS_ORIGINS is not set
- Support multiple origins by splitting and stripping whitespace from env var
- Validate filename ends with .csv (case-insensitive)
- Validate MIME type is text/* or application/csv or text/csv
- Return HTTP 400 with error message if validation fails
- Mark task 2.4 as complete
- Reject uploads larger than 10MB before reading file content
- Reject CSVs with more than 500,000 data rows after parsing
- Checks placed as early as possible in the handler flow
- Mark task 2.3 as done in tasks.md
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add `import re` to services/ml/app/main.py
- In POST /model/load: validate run_id matches ^[a-zA-Z0-9_-]+$ before DB lookup; use Path.resolve() + directory containment check before loading model artifact
- In DELETE /training/runs/{run_id}: validate run_id matches ^[a-zA-Z0-9_-]+$ before any processing; use Path.resolve() + directory containment check before deleting model artifact
- Both endpoints return HTTP 400 with {"detail": "Invalid run_id format"} on invalid input
- Mark task 2.2 as completed in openspec/changes/code-review-fix/tasks.md
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Validate that run_id matches /^[a-zA-Z0-9_-]+$ regex before interpolating into the API URL.
Returns HTTP 400 with 'Invalid run_id format' error if validation fails.
This prevents potential URL injection attacks and invalid identifier usage.
Changes:
- Updated docker-compose.yml MLflow service port binding from 5000:5000 to 127.0.0.1:5000:5000
to restrict access to localhost only for security
- Marked task 1.7 as complete in tasks.md
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Changed PostgreSQL service port binding from 5432:5432 to 127.0.0.1:5432:5432 in docker-compose.yml
- This restricts PostgreSQL to listen only on localhost, improving security by preventing access from other interfaces
- Marked task 1.6 as completed
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
All DATABASE_URL values and postgres service env vars now use
\${POSTGRES_USER}, \${POSTGRES_PASSWORD}, \${POSTGRES_DB} interpolation
instead of hardcoded ml_user/ml_password/candle_annotator values.
Also updated pg_isready healthcheck to use the same env vars.
Closes task 1.5.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Remove hardcoded SQL comments containing 'ml_user' and 'ml_password'
- Remove fallback default credentials in DATABASE_URL construction
- Add fail-fast validation: raise RuntimeError if DATABASE_URL env var is missing or empty
- Mark task 1.4 as complete in code-review-fix/tasks.md
- Replace ml_password with change_me_to_a_strong_password placeholder
- Replace ml_user with your_db_user placeholder
- Mark task 1.3 as completed in tasks.md
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Add build_dataset_from_db() that exports candles from DB, runs feature
engineering, and ingests span annotations into labeled CSV
- Call it automatically in _run_training_background before training starts
- Add POST /training/build-dataset endpoint for standalone use
- Add Next.js proxy route /api/training/build-dataset
- Update TrainingPanel: remove dataset-missing block on Start Training,
show informational message that dataset builds automatically
- Rewrite scripts/run-migrations.js for PostgreSQL (was better-sqlite3)
- Rewrite scripts/load-initial-data.js for PostgreSQL (was better-sqlite3)
- Make db connection lazy in src/lib/db/index.ts to avoid build-time errors
when DATABASE_URL is not available in Docker build stage
