知识库

I need to initialize resources (database connection pool, Redis client, HTTP client) when my FastAPI app starts and clean them up when it shuts down. The old @app.on_event('startup') pattern is deprec...

I want to use FastAPI's dependency injection system cleanly across route handlers. I have database sessions, current user extraction from JWT, and Redis clients that need to be injected. I want to use...

After handling an API request (e.g., creating a trace), I want to trigger background processing (generate embeddings, send a notification email) without blocking the response. I need this to run after...

I need to manage configuration for a FastAPI app across local dev, CI, and production environments. I use environment variables and .env files. I want type-safe settings with validation, defaults, and...

I need to bulk insert thousands of rows into PostgreSQL using SQLAlchemy 2.0 async. I want to do it efficiently with a single query (not one INSERT per row) and get back the generated IDs without a se...

I'm getting 'MissingGreenlet' or 'greenlet_spawn has not been called' errors when accessing related objects on my SQLAlchemy models in an async context. I need to understand which lazy loading strateg...

I need to version my FastAPI API (v1, v2 eventually) without breaking existing clients when I introduce breaking changes. I want to organize routes in separate router modules per version and apply ver...

I need structured JSON logging in my FastAPI application. I want request IDs correlated across log lines, log levels configurable via env var, and logs formatted as JSON in production but human-readab...

I need to run multiple async operations concurrently in Python and collect all their results. Some operations may fail and I want to handle errors per-task. I'm using Python 3.11+ and want to use the ...

I need to write async pytest tests for my FastAPI application. I need a test database that's rolled back after each test (not truncated), and I need to override FastAPI's database dependency so tests ...

I want to measure API request duration and expose it as a Prometheus metric. I need middleware that times every request, records the route, method, and status code, and exposes a /metrics endpoint for...

I need to implement a paginated list endpoint in FastAPI. The endpoint accepts optional filters (status, tag name) and pagination parameters (page, page_size). I want to avoid N+1 queries and get tota...

My API uses snake_case internally (Python convention) but the frontend expects camelCase JSON. I need Pydantic models that accept and output camelCase JSON while keeping snake_case attribute names in ...

I have a function that handles multiple edge cases and I want to test all of them without writing a separate test function per case. I want to use pytest's parametrize decorator to test with different...

I need to rate limit my FastAPI API endpoints. I want different limits for read vs write operations (e.g., 60 reads/min, 20 writes/min per API key). The rate limiter must be atomic to handle concurren...

I have a background worker (not FastAPI) that processes traces in a loop. Each batch of traces needs its own database session. I need to avoid session reuse across batches, handle connection pool exha...

My FastAPI application raises various exceptions in different layers (validation errors, database errors, business logic errors) and I want all errors to return the same JSON structure. I also want to...

I need to make HTTP calls to external APIs from my FastAPI application. I want a shared client with connection pooling (not a new client per request), configurable timeouts, and proper error handling....

I need to create an async context manager for managing resources that need cleanup (database connections, temp files, locks). I want to use it with both `async with` syntax and as a decorator.

I have a SQLAlchemy ORM model with sensitive fields (api_key_hash, internal_flags) that I don't want to expose in API responses. I also want to add computed fields to the response (e.g., is_verified c...

I need to enforce a composite unique constraint across multiple columns in SQLAlchemy (e.g., one vote per user per trace, unique domain reputation per user per tag). I also need a partial unique index...

I have a background worker that processes items concurrently, but I don't want to overwhelm the database or external API. I need to limit the number of concurrent coroutines processing at any time.

I need to add a new column to an existing PostgreSQL table that already has data. The column has a default value. I want to avoid locking the table and need to understand the migration ordering to han...

I need to push real-time updates to browser clients when traces are validated or new search results arrive. I want to use WebSockets with FastAPI, handle connection cleanup when clients disconnect, an...

I'm building a Python service and not sure whether to use Python dataclasses, Pydantic models, or attrs for different data structures. I need to understand the tradeoffs for API schemas, internal data...

My application makes calls to the OpenAI API and other external services. I want to write unit tests that don't make real HTTP calls, can test different response scenarios (success, error, timeout), a...

I'm writing Python code with strict type checking (mypy) and need to properly annotate functions that work with generic types: functions returning different types based on input, TypeVar with bounds, ...

My API has endpoints that return large datasets (100k+ rows). Offset-based pagination becomes slow at high page numbers. I need to implement cursor-based pagination that's efficient at any position in...

I need to insert records but update them if they already exist (upsert). This is common for incrementing counters, updating reputation scores, or syncing external data. I need the PostgreSQL-specific ...

I want to customize FastAPI's automatically generated OpenAPI documentation: add authentication scheme to Swagger UI, add custom descriptions and examples to request/response models, group endpoints b...

I need to stream large query results from PostgreSQL without loading everything into memory. I want to use a Python async generator that yields rows in batches, and I want to support both streaming HT...

My pytest test suite is growing and I'm having trouble managing shared fixtures. I need to organize conftest.py files across a test hierarchy, understand fixture scoping, and share fixtures between di...

I need a /health endpoint for Docker healthchecks and load balancer probes. The endpoint should verify the database connection and Redis connection are alive, return structured health status, and resp...

I need to automatically track when rows are inserted or updated in my database tables for audit purposes. I want to log changes without modifying every query in my codebase.

I'm using Python Enum classes for status fields and vote types in my application. I need them to work correctly with SQLAlchemy (stored as strings in PostgreSQL, not as database enum types), with Pyda...

I need an endpoint that accepts file uploads (CSV, JSON, images). I need to validate file type, size limit the upload, and process the file asynchronously without blocking the server. I want to handle...

I have a slow PostgreSQL query and I ran EXPLAIN ANALYZE. I need to understand how to read the output, identify why the query is slow (wrong index, bad join order, row estimate mismatch), and know wha...

My FastAPI application has 20 workers and each worker has a SQLAlchemy connection pool of 5. This creates 100 connections to PostgreSQL, which is hitting the max_connections limit. I need connection p...

I have a PostgreSQL table where most queries filter on a specific condition (e.g., status='validated'). Creating a full index wastes space on non-matching rows and slows down writes. I want to create ...

I have multiple API instances behind a load balancer and need to broadcast events (trace validated, new search result) to all instances. Redis pub/sub will fan out messages to all subscribers. I need ...

My application has concurrent requests that update the same rows in different orders, causing deadlocks (ERROR: deadlock detected). I need to understand why deadlocks happen and how to prevent them in...

My Docker image for a FastAPI app is 2GB and takes 5 minutes to build. I need a multi-stage Dockerfile producing a slim production image, using layer caching effectively to avoid reinstalling packages...

My Docker Compose setup has FastAPI, PostgreSQL, Redis, and a background worker. The services need to talk to each other. I am confused about when to use service names vs localhost, and how ports work...

I am deploying FastAPI with Nginx as a reverse proxy. I need SSL termination, proper forwarding headers (X-Forwarded-For), static file serving, and connection keepalive to the upstream FastAPI app.

I need to manage environment variables for Docker Compose across multiple environments without committing secrets. I want .env file support with per-environment overrides.

I keep running into React hooks issues: state not updating immediately after setState, stale values in event handlers, and batching behavior I do not understand.

My React components have useEffect hooks causing memory leaks, firing too often from missing dependencies, or not re-running when I change a filter object. I need to understand cleanup and dependency ...

I have API call state with loading, success, and error variants. Using optional fields allows impossible states. I want discriminated unions with full TypeScript narrowing.

I have full TypeScript types for my API responses and need derived types: making all fields optional for updates, picking a subset for list views, omitting sensitive fields, and requiring specific fie...

I need to share authenticated user state across many components without prop drilling. I want React Context with a pattern that avoids unnecessary re-renders when the setter is called but the user dat...

I am building with Next.js App Router and confused about when to use server vs client components, how to fetch data in server components, and how mutations work with server actions.

I need GitHub Actions to deploy to staging and production with secrets (API keys, deploy tokens) scoped per environment, with production requiring manual approval before deploy.

I want to deploy my FastAPI application to fly.io with a managed PostgreSQL database. I need fly.toml configuration, migration execution before deployment, and secure secret management.

I need free HTTPS for my production server. I am using Nginx as a reverse proxy and need automatic certificate renewal with Let's Encrypt.

I have PostgreSQL with pgvector embeddings and a trust_score. I want semantic search that combines vector similarity with trust score for final ranking, without cutting off high-trust results before r...

I need to rank traces by user votes. Simple upvote/total ratio fails for low-vote items (1/1 = 100% looks better than 95/100 = 95%). I need Wilson score lower bound for confidence-interval-aware ranki...

I need full-text search in PostgreSQL for finding traces by keywords. LIKE queries are too slow and do not support stemming. I want indexed text search with relevance ranking.

I have multiple API instances behind a load balancer and need to broadcast events (trace validated, vote cast) to all instances simultaneously. Redis pub/sub fans out to all subscribers.

I have Redis and a Python background worker in Docker Compose. I need healthchecks so dependent services only start once Redis is ready, and I want the worker to report healthy only when it is activel...

I need to retry failing async operations (external API calls, database operations under transient load) with exponential backoff. I want a reusable decorator and to retry only on specific exception ty...

I need to add a new NOT NULL column to an existing PostgreSQL table that has data. I need to avoid table locking in production and understand the correct staging for nullable -> backfill -> NOT NULL.

I need fast async pytest tests for my FastAPI application using a real database. I want tests to roll back after each test rather than truncating tables, and override the FastAPI database dependency w...

I am building forms in React that have re-render performance issues from controlled inputs, complex validation, and async submission errors from the API. I want a clean performant solution.

I want to stream OpenAI chat completion responses to users so tokens appear in real-time. I need to handle streaming in the backend (FastAPI) and forward it to the frontend using Server-Sent Events.

I need to integrate Stripe Checkout for subscription payments. Users click a button, get redirected to Stripe's hosted checkout page, complete payment, and return to my app.

My React app re-renders excessively. I want practical guidance on when useMemo and useCallback actually help vs when they are premature optimization that adds overhead without benefit.

I am migrating from Pydantic v1 to v2 and need custom validation: field-level validators that transform input (normalize strings, coerce types), cross-field validators comparing multiple fields, and c...

I want to measure API request duration and expose Prometheus metrics. I need middleware that times every request, records route and status code, and serves a /metrics endpoint for scraping.

I need a paginated list endpoint in FastAPI with optional filters (status, tag). I want efficient pagination with total count and no N+1 queries.

I need to understand best practices for async JavaScript/TypeScript: when to use Promise.all vs Promise.allSettled, how to handle errors in async patterns, and how to avoid common pitfalls like unhand...

My FastAPI application raises various exceptions in different layers and I want all errors to return a consistent JSON structure. I also want to map internal exceptions to appropriate HTTP status code...

I have PostgreSQL queries that always filter on status='validated' and most queries only target this subset. A full index wastes space indexing pending rows I never query. I want partial indexes for f...

My API returns large datasets and offset-based pagination becomes slow at high page numbers. I need cursor-based pagination that stays O(log n) at any position in the dataset.

I need to decide when to use Python dataclasses, Pydantic models, TypedDict, or NamedTuple for different data structures in my application. Each has tradeoffs in validation, serialization, and overhea...

I need to bulk insert thousands of rows into PostgreSQL efficiently. Individual inserts in a loop are too slow. I want a single bulk INSERT and need the generated IDs back without a second SELECT.

I need to run multiple async operations concurrently and collect all results. I want to handle cases where some may fail independently and need Python 3.11+ structured concurrency approach.

I have a function handling many edge cases and want to test all of them without a separate test function per case. I want pytest parametrize with multiple inputs and expected outputs including error c...

I need to version my FastAPI API (v1, v2 eventually) without breaking existing clients. I want routes organized in separate modules per version with versioned URL prefixes and shared authentication.

I keep getting MissingGreenlet errors or N+1 queries when accessing related objects on SQLAlchemy models in an async context. I need to understand which loading strategies work with async and how to e...

My React app has a large JavaScript bundle. I want to split the code so users only download JS for the current page using React.lazy for component-level code splitting.

I need to push real-time updates (trace validated, vote cast) to browser clients. I want WebSockets with FastAPI, clean handling of client disconnections, and broadcast to all connected clients.

I need to automatically track when rows are inserted or updated in PostgreSQL for audit purposes. I want this without modifying every query in my codebase.

After handling an API request (creating a trace), I want to trigger background processing (generate embeddings) without blocking the response. The work should run after the response is sent to the cli...

I need to manage configuration for a FastAPI app across local dev, CI, and production environments using environment variables and .env files. I want type-safe settings with validation and defaults.

I have a SQLAlchemy ORM model with sensitive fields I don't want in API responses (api_key_hash, internal flags). I also want computed fields in the response. I need Pydantic v2 ORM mode.

I am using Python Enum classes for status fields in my application and need them to work with SQLAlchemy (stored as strings), with Pydantic validation, and with JSON serialization without a custom enc...

I need to unit test code that depends on external services (OpenAI, database, Redis). I want to mock these dependencies to test my business logic in isolation without real I/O.

I have concurrent reads and writes and am seeing unexpected behavior: dirty reads, non-repeatable reads, or phantom rows. I need to understand PostgreSQL isolation levels and when to use each.

I have trace metadata that varies by domain (language version, framework name, OS). I don't want to add columns for every possible field. I need JSONB storage with indexing for specific keys.

I have expensive database queries for search results I want to cache in Redis. I need a clean cache-aside pattern that serializes Pydantic models, handles Redis unavailability gracefully, and avoids s...

I have a slow PostgreSQL query and ran EXPLAIN ANALYZE but don't know how to read the output. I need to identify why the query is slow (wrong index, bad join order, row estimate mismatch) and know wha...

I want to use a base docker-compose.yml for common configuration and override files for environment-specific settings (development with hot reload, production with resource limits, CI with test setup)...

My Docker builds are slow because every code change invalidates the package installation layer. I need to structure my Dockerfile so that dependency installation is cached and only code changes trigge...

I need reusable data fetching hooks for my React application that handle loading, error, and success states, cache responses, revalidate on focus, and support optimistic updates.

I enabled TypeScript strict mode and now have many type errors. I need to understand what strict mode enables, how to fix common errors (possibly undefined, implicit any), and how to migrate existing ...

I want GitHub Actions to automatically create version tags and GitHub Releases when I push to main. I use conventional commits (feat:, fix:, chore:) and want semantic version bumps based on commit typ...

My GitHub Actions workflows take 10+ minutes mainly due to installing dependencies on every run. I need to cache dependencies, understand what to cache for Python and Node.js, and handle cache invalid...

My GitHub Actions workflows queue up multiple runs when I push rapidly. I want to cancel old runs when a new commit is pushed to the same branch, while still running all checks on main.

I am hitting OpenAI rate limits and getting RateLimitError exceptions. I need robust retry logic with exponential backoff specifically for OpenAI API calls, and I need to handle different error types ...

I need to handle webhooks from multiple providers (Stripe, GitHub, custom services). I need a reusable pattern for signature verification, idempotent processing, and handling duplicate deliveries.

My tests need realistic data (traces with tags, users with votes). Creating data manually in each test is verbose and fragile. I want factory fixtures that generate realistic test data with sensible d...

I want to go beyond example-based tests to automatically discover edge cases in my code. I need property-based testing that generates diverse inputs and finds cases I would not think of manually.

I need consistent structured logging across my Python application. Logs need to be machine-parseable JSON in production but human-readable in development, with context (request_id, user_id) propagated...

I need API key authentication for my FastAPI service. API keys must be stored securely (not plaintext), validated on every request, and I want to support multiple keys per user with revocation.

I need to stream large data exports from my FastAPI application. I want to generate NDJSON (newline-delimited JSON) as an async generator and stream it to the client without loading all data into memo...

I have expensive aggregation queries (contributor statistics, tag popularity, trust score distributions) that run on every page load. I want to precompute these and refresh them periodically.

I have multiple API instances and need to ensure only one runs a specific operation at a time (e.g., leaderboard rebuild, scheduled job). I need a distributed mutex without adding a separate locking s...

Alembic's autogenerate misses some schema changes: custom PostgreSQL types, check constraints, and index changes. I need to configure autogenerate to detect more changes and add manual DDL for unsuppo...

I want TypeScript to check that objects match a type while preserving the most specific (narrowest) type for inference. I also want const objects with literal types not broadened to string.

My modal component is inside a div with overflow:hidden or z-index stacking context that clips it. I need to render the modal at the document.body level while keeping it controlled by my React compone...

I need to extend types from a third-party library (fastify, express, next.js) to add my own properties (current user, request context) without modifying the library's source code.

I need my GitHub Actions workflows to access AWS services (S3, ECR, ECS) without storing long-lived AWS credentials as GitHub secrets. I want keyless authentication using OIDC.

I have multiple FastAPI instances behind a load balancer. Simple in-memory rate limiting does not work because requests are distributed across instances. I need Redis-based rate limiting that works ac...

I want to let users log in with their GitHub account. I need to handle the OAuth2 flow: redirect to GitHub, receive the callback code, exchange for user info, and create/update a user in my database.

I need end-to-end tests that verify my web application works from the user's perspective. I want to test user flows (search, vote, contribute) in a real browser using Playwright.

I want to measure test coverage in my Python project, enforce minimum coverage thresholds in CI, and track coverage changes over time. I use pytest and want coverage reported in the CI run.

I have Python code using os.path and string concatenation for file paths that breaks on Windows. I need to use pathlib for cross-platform file operations that work on Linux, macOS, and Windows.

I need to build a command-line interface for my application with subcommands (import, export, stats), options with validation, and help text. I want it to be testable and support both interactive and ...

I am using Python dataclasses for internal data objects. I need mutable default values (lists, dicts), computed fields that derive from other fields, and validation in __post_init__.

I have a many-to-many relationship between traces and tags. I need to create the join table correctly in SQLAlchemy, handle tag insertion with get-or-create semantics, and query traces by tag efficien...

I need my FastAPI routes to participate in database transactions where multiple operations must succeed or fail together. I want a dependency that manages the transaction lifecycle.

I have a complex UI component (a card with header, body, footer, and actions) that is hard to reuse because the structure is too rigid. I want a flexible composition pattern that lets callers customiz...

I am building a multi-tenant SaaS where each user should only see their own data. I want PostgreSQL row-level security (RLS) policies to enforce data isolation at the database level.

Need to add full-text search to a PostgreSQL table. Users type natural-language queries like 'python async error handling' and expect relevant rows returned quickly without external search infrastruct...

Need to enforce data isolation so each tenant can only see their own rows. Application-level filtering is error-prone, a missing WHERE clause leaks all data. Want database-enforced isolation.

Storing metadata as JSONB columns for flexibility, but queries like `WHERE data->>'status' = 'active'` do full table scans. Need to index JSONB fields and write efficient queries.

Need to push real-time updates to connected clients when events occur (e.g., trace validated, vote received). Using Server-Sent Events (SSE) for the client side, need Redis pub/sub to fan out events f...

Need a real-time leaderboard that updates as scores change. Querying PostgreSQL for top-N with ORDER BY is expensive at scale. Need O(log N) updates and O(log N + K) range queries.

Managing different configurations for development, staging, and production in Docker Compose. Hardcoding environment variables in compose files leads to secrets in version control and different config...

Python Docker images are large (1GB+) because they include build tools, pip cache, and development packages. Need a lean production image while keeping a full dev environment. Using uv for fast depend...

FastAPI is running behind Nginx. Need rate limiting per client IP to prevent abuse, response caching for expensive endpoints, and gzip compression. Currently serving all traffic directly without any o...

Docker containers are ephemeral, data stored inside the container is lost on restart. Need to persist PostgreSQL data, handle Redis persistence, and share files between containers. Confusion between n...

Component state has grown complex with multiple related fields that change together. useState is getting unwieldy with many separate setters. Need predictable state transitions and easier debugging.

Duplicating data-fetching logic across components, each API call has its own loading/error/data state management. Need a generic, reusable data-fetching hook that works with any endpoint and return ty...

Using Next.js App Router. Need to handle form submissions that mutate data on the server without building a separate API route. Want type-safe server-side form handling with optimistic updates.

API calls return different shapes depending on success or failure. Using a generic `{ data: T | null, error: string | null }` pattern leads to redundant null checks everywhere and TypeScript can't nar...

Multiple deeply nested components need access to the same state (current user, theme, feature flags). Prop drilling has become unmanageable. Want to avoid Redux for a smaller app but need predictable ...

Processing a large CSV or JSON file (hundreds of MB) by reading the entire file into memory causes out-of-memory errors. Need to process records one at a time as they're read from disk.

Need to build a Docker image and push it to GitHub Container Registry (GHCR) on every push to main, with proper tagging (latest, sha, and version tags). Want to avoid rebuilding unchanged layers.

Multiple repositories have duplicate CI/CD logic (lint, test, build). Updating the workflow in each repo separately is error-prone. Need a single source of truth for shared CI steps.

CI pipeline runs tests automatically, but production deployment should require manual approval. Want automated staging deploy on merge to main, but production needs a human sign-off before deploying.

Using OpenAI chat completions but responses have high latency before any text appears. Need to stream the response token-by-token so users see text as it's generated, rather than waiting for the full ...

Implementing Stripe webhooks. Stripe sends events when payments succeed, subscriptions change, or invoices are created. Need to verify webhook signatures to prevent spoofed events and handle retries i...

Need to send transactional emails (welcome emails, password resets, notifications) from a FastAPI application. Evaluating Resend as a simpler alternative to SendGrid/Mailgun with a clean Python SDK.

Writing tests for a function that should behave differently based on many input combinations. Duplicating test functions for each case leads to hundreds of lines of copy-pasted code. Need a clean way ...

Unit tests mock too much and don't catch integration bugs between routes, middleware, and database. Need integration tests that test the full HTTP stack (auth, rate limiting, actual DB queries) withou...

Tests call external APIs (OpenAI, Stripe, GitHub). Real API calls make tests slow, expensive, and flaky. Need to intercept HTTP calls at the transport layer so the application code doesn't need to cha...

Async pytest fixtures are slow because they recreate expensive resources (database connections, HTTP clients) for each test. Need to understand fixture scoping in async contexts and how to share resou...

Using Pydantic BaseModel for everything including internal data transfer objects (DTOs) that never touch the API boundary. Pydantic validation overhead is unnecessary for internal models. Need guidanc...

Executing multiple independent async operations sequentially instead of concurrently. Using `asyncio.gather()` but need better error handling when one task fails, gather continues other tasks even on ...

SQLAlchemy ORM queries are unexpectedly slow. EXPLAIN ANALYZE shows hundreds of small queries instead of a few efficient ones. The N+1 problem: loading a list of traces then accessing trace.tags trigg...

Expensive computations (tag normalization regex compilation, settings lookups, configuration parsing) run repeatedly with the same inputs. Need simple memoization without external caching infrastructu...

FastAPI services are tightly coupled to concrete implementations (PostgreSQL, Redis, specific email provider). Want to swap implementations for testing without monkey-patching or complex mocking setup...

Need to split a single `name` column into `first_name` and `last_name` columns. This requires a schema change (add columns, remove old) AND a data migration (populate new columns from old). Pure DDL m...

Building code that needs to work with both asyncio and Trio event loops (or library code that shouldn't dictate the event loop). Also need nursery-style task cancellation that TaskGroup provides but w...

Application raises various exceptions (ValidationError, NotFound, PermissionDenied) but they all return inconsistent error responses. Clients can't reliably parse error details. Need a consistent erro...

Getting errors like 'There is no current event loop' or 'coroutine was never awaited' in async Python code. Also confusion about when to use asyncio.run() vs await, and how to call async code from syn...

Queries filter on multiple columns: `WHERE status = 'validated' AND trust_score > 0.5 ORDER BY created_at DESC`. Single-column indexes on each column aren't used effectively. Query planner does a full...

Pydantic models need computed/derived fields that are calculated from other fields (e.g., full_name from first/last, formatted dates, masked API keys). Also need custom serialization (snake_case to ca...

Docker Compose healthchecks fail because the container doesn't have curl or wget installed in a slim image. Need a working healthcheck that works in minimal Alpine and Debian-slim images, and handles ...

Manually managing server state with useState/useEffect is complex: tracking loading/error/stale states, deduplicating requests, invalidating cache after mutations. Looking for a dedicated server state...

Using OpenAI chat completions to extract structured data from user input (classify intent, extract entities, fill forms). Parsing JSON from unstructured LLM output is fragile and requires complex prom...

Query is slow but unsure why. Running EXPLAIN ANALYZE produces verbose output with nodes like 'Seq Scan', 'Hash Join', 'Bitmap Heap Scan' and numbers for cost, rows, and buffers. Need to understand wh...

Multiple workers process tasks concurrently and need to ensure only one worker handles a given resource at a time. Need a distributed lock that works across multiple API instances with automatic expir...

Inserting thousands of rows one at a time with individual session.add() calls is too slow. Need bulk inserts that return the auto-generated IDs of inserted rows for downstream processing.

Docker Compose file has many services but not all are needed all the time. Dev needs API + DB, testing needs test DB too, production needs different services. Maintaining separate compose files leads ...

Docker images need different configuration for different environments (staging vs production). Hardcoding config in the Dockerfile or passing everything at runtime isn't sufficient, some values must b...

Need to protect certain routes in Next.js App Router, redirect unauthenticated users to /login, redirect logged-in users away from /login to dashboard. Doing this in each page component leads to flash...

Defining configuration objects or lookup maps with TypeScript. Using `as const` loses type checking, using explicit type annotations loses inference of literal types. Need both: type checking AND infe...

Bundle size is large because all components load upfront. Some pages are rarely visited. Need to split the bundle so users only download code for the routes they visit.

CI/CD pipeline needs secrets (API keys, deployment credentials) for different environments. Using repository-level secrets means production keys are accessible in all workflows including untrusted PRs...

Deploying a FastAPI application to Fly.io. Need to configure machine sizes, auto-scaling, health checks, persistent volumes for file storage, and secrets management for the deployment.

Calling the GitHub REST API to fetch repository data, user information, and pull request details. Hitting rate limits (60 req/hour unauthenticated, 5000/hour with token). Need to handle pagination and...

Need to store and serve user-uploaded files (images, PDFs). Storing files in the filesystem doesn't work with multiple API instances or ephemeral containers. Need object storage that works with AWS S3...

Integration tests need a real PostgreSQL database but spinning up a persistent test database causes state pollution between test runs, requires manual setup, and breaks in CI without a running Postgre...

Working with union types and need TypeScript-style type guards in Python. After checking `isinstance(x, str)`, the type checker should know `x` is a `str` in that branch. Also need custom narrowing fu...

Test fixtures are duplicated across multiple test files. Each file has its own database setup, mock clients, and test data factories. Need a way to share fixtures across an entire test suite without i...

Resources like database connections, file handles, locks, and HTTP clients need cleanup even when exceptions occur. Using try/finally everywhere is verbose. Need the context manager pattern for both c...

Need real-time bidirectional communication between browser clients and the FastAPI server. REST/SSE only support server-to-client. Building a live collaboration feature where multiple users see update...

Using string literals for status values ('pending', 'validated', 'active') spread across the codebase. Typos cause silent bugs, IDEs can't autocomplete, and it's hard to find all places a status value...

Using Python's standard logging module but log output is unstructured text that's hard to search in log aggregation tools (Datadog, Grafana Loki, CloudWatch). Need structured JSON logs with consistent...

Storing OpenAI embedding vectors in PostgreSQL with pgvector. Need to query for the N most semantically similar traces given a query embedding. Query must filter by tags and status before vector ranki...

Need to track request latency, status code distribution, and active request counts across all endpoints. Want Prometheus metrics that work with Grafana dashboards, without duplicating instrumentation ...

Application needs configuration from multiple sources: environment variables override .env file, which overrides defaults. Also need nested settings objects (database config, Redis config) and type co...

Deleting a parent record (User) should automatically delete child records (Traces, Votes). Without proper cascade configuration, SQLAlchemy either raises a foreign key constraint error or leaves orpha...

Building a type-safe event system where events are named 'resource:action' (e.g., 'trace:created', 'user:updated'). Need TypeScript to enforce valid event names and infer payload types from event name...

Background tasks are blocking the API event loop, image processing, email sending, report generation take seconds. FastAPI BackgroundTasks run in the same event loop. Need a proper job queue that runs...

FastAPI application runs behind Nginx as a reverse proxy. Regular HTTP requests work but WebSocket connections fail with 400 or 502 errors. Need Nginx configured to properly proxy WebSocket upgrade re...

Building a complex form with nested fields, cross-field validation, and async validation (e.g., check if username is taken). Using controlled components with useState for each field is cumbersome and ...

Multiple service implementations (email: Resend, SendGrid; storage: S3, R2, local) need to be swappable. Using duck typing alone makes it unclear what methods a service must implement. Need formal int...

Blog posts and documentation pages are generated server-side on every request (SSR), causing slow TTFB. Content changes infrequently. Need static generation benefits (CDN caching, fast delivery) with ...

FastAPI application with SQLAlchemy has database connection issues under load: timeouts, 'too many connections' PostgreSQL errors, and connection leaks. Default connection pool settings are not tuned ...

Monorepo contains multiple services (api/, frontend/, mcp-server/). Every push runs all CI pipelines even when only one service changed. Need to run only the relevant pipelines based on which files ch...

FastAPI database dependency using SQLAlchemy async session. Need automatic transaction rollback when route handlers raise exceptions, and ensure sessions are always closed without leaking connections....

Calling external APIs (OpenAI, Stripe, GitHub) that occasionally return 429 (rate limit) or 503 (transient errors). Need automatic retry with exponential backoff, jitter to avoid thundering herd, and ...

Alembic migration needs to modify an existing column (change type, add constraint) in a way that works on both PostgreSQL (production) and SQLite (tests/development). PostgreSQL supports ALTER COLUMN ...

Need to run Alembic migrations that work on both PostgreSQL and SQLite. SQLite doesn't support ALTER TABLE for column renames or type changes, causing migration failures in test environments.

Need to implement rate limiting for an API. Token bucket is too bursty, and fixed window has the boundary problem where a user can double their limit across two windows.

Building a typed API client where response types differ based on request parameters. Need type-safe response handling without runtime checks or type assertions.

CI pipeline uses matrix builds for multiple Python versions but stops all jobs when one fails. Need all matrix combinations to complete so developers see the full picture of compatibility.

MCP server (FastMCP 3.0.0) deployed to Railway crashes at runtime with `ModuleNotFoundError: No module named 'httpx'`, even though `httpx>=0.27` is listed under `[project].dependencies` in pyproject.t...

In a Python repo with `pyproject.toml` dependency groups and `uv.lock`, direct `pytest`, `python -m pytest`, and `python3 -m pytest` may fail (`command not found` or `No module named pytest`) dependin...

Running `uv add <package-name> --editable <path>` fails with a misleading error: "<package-name> did not resolve to a local directory, but the --editable flag was provided." uv treats the first positi...

During an OAuth service migration, static type checks passed after adding compatibility shims, but unit tests still failed because previously-awaited helper methods and strategy hooks were removed or ...

Pydantic v2 `bytes` fields serialize text-like content to JSON fine but raise `PydanticSerializationError: invalid utf-8 sequence` for actual binary data. This surfaces only when `model_dump_json()` i...

As of 2026, the astral-sh/setup-uv GitHub Action stopped publishing major-only version tags (e.g., @v8) for security reasons. Using @v8 will fail to resolve. This is a departure from the common GitHub...

`from pptx import Presentation` returns a factory function, not the class. The actual class at `pptx.presentation.Presentation` has a docstring explicitly stating "Not intended to be constructed direc...

Python's site-packages mechanism executes any line starting with `import` in a *.pth file at interpreter startup, before user code runs. The compromised litellm 1.82.7/1.82.8 (PyPI, ~40 min live, 2026...

In SQLModel 0.0.x, `Session.exec()` wraps SQLModel statements and does not accept the SQLAlchemy-style `(statement, params)` call shape. Using `Session.exec(text_stmt, params)` raises `TypeError: Sess...

Base64Bytes validates input by attempting to base64-decode it. Passing raw binary bytes (e.g. b"\x89PNG...") does not raise a validation error, Pydantic interprets the raw bytes as base64 characters a...

In OpenTelemetry Python setup code, using one `_initialized` flag for both "setup ran" and "SDK enabled" can cause disabled mode bugs: setup returns early for `OTEL_SDK_DISABLED=true`, but later app i...

PyJWT 2.10+ added InsecureKeyLengthWarning in jwt/api_jwt.py that fires when the HMAC key used with HS256 is shorter than 32 bytes (256-bit hash output), per RFC 7518 Section 3.2. The check is byte-le...

apron-auth 0.2.0 introduces OAuthClient.fetch_identity with built-in handler inference for Google and GitHub based on provider endpoint hostnames (including manually-constructed provider configs targe...

When wrapping callables defined in a module that uses ``from __future__ import annotations`` (PEP 563), ``inspect.signature(fn).parameters[name].annotation`` returns the annotation as a *string* (the ...

When upgrading or re-installing ty (Astral's type checker), versions from 2026-04-28 onward (0.0.33+) flag valid pydantic model constructions where the python attribute name differs from the alias as ...

When using Hatchling, the wheel builder tries heuristics to select files to ship. If your distribution name (e.g. `craic-demo`) implies a normalized module name (e.g. `craic_demo`) but your repo doesn...

Some OAuth providers (e.g. HubSpot's `/oauth/v1/access-tokens/{access_token}` introspection endpoint) carry the bearer token in the URL path rather than an Authorization header. When the request fails...

SQLAlchemy's DateTime(timezone=True) column type is a hint that PostgreSQL honors but SQLite silently ignores. Values stored via datetime.now(UTC) are persisted without timezone info in SQLite and ret...

Starting with openai SDK 2.34.0 (released around 2026-05-11), passing api_key="" or api_key=None to AsyncOpenAI() raises openai.OpenAIError: Missing credentials. Previous versions (2.33.0 and earlier)...

The detect-secrets pre-commit hook scans staged diffs, not just newly-introduced lines. If you edit a line near an existing test fixture containing a literal that looks secret-like (e.g. a fake `passw...

Verified in a Python project using `[tool.uv.sources]` with `git = "https://..."` and `branch = "main"`: resolved packages in `uv.lock` were pinned to explicit commit SHAs (`...#<commit>`), and instal...

The OpenAI Python SDK's embeddings.create() method automatically sets encoding_format="base64" when the parameter is not explicitly provided, then decodes the base64 response client-side for performan...

In optimistic-concurrency flows that use conditional UpdateItem for one-time token/state consumption, a subsequent default (eventually consistent) GetItem can lag and misclassify failures under conten...

In a release workflow for a `schema/` submodule, the tag validator accepted `schema/X.Y.Z` but rejected `schema/vX.Y.Z`. Go module tags for subdirectory modules are typically `subdir/vX.Y.Z`, so rejec...

In a pnpm-managed frontend, we observed a corrupted modules layout where `node_modules/.pnpm` existed and package tarballs were present, but no top-level symlinks, no `.bin`, and no `.modules.yaml`. T...

Common test fixture pattern: at setup, snapshot the ids of pre-existing rows in a table (the seeded ones we want to survive across modules). At teardown, delete every row whose id is NOT in that snaps...

When a service is pure with respect to transactions, it does not call session.commit() or session.rollback() itself, leaving lifecycle management to the framework's session dependency (e.g. FastAPI's ...

When installing in-memory localStorage/sessionStorage shims in a Vitest setupFile under `environment: 'jsdom'`, the `window` object IS the global object, so `Object.defineProperty(globalThis, 'localSt...

To simulate a GitHub release event with nektos/act, create a JSON file with the event payload (e.g. {"action":"published","release":{"tag_name":"v1.0.0"}}) and pass it via -e. For workflow_dispatch, u...

Google API IDs (spreadsheet IDs, file IDs) are long base64-like strings that trigger `detect-secrets`'s Base64HighEntropyString detector. This blocks commits containing realistic testdata JSON fixture...

A module-scoped conftest db fixture that snapshots pre-existing IDs in pivot tables (e.g. User, Namespace) and on teardown deletes rows whose IDs aren't in the snapshot works only while every test-cre...

A common pattern: module-scoped db fixture snapshots existing row IDs at setup and deletes only non-snapshot rows at teardown. Two failure modes: (a) rows seeded by a now-removed init path persist ind...

Pydantic v2 BaseSettings stores field values on instances, not the class. unittest.mock.patch.object(type(settings), "FIELD", value) raises `AttributeError: <class 'Settings'> does not have the attrib...

In a React+Vitest+jsdom frontend suite, many tests failed with `localStorage.* is not a function` while Node emitted `--localstorage-file was provided without a valid path`. A robust fix was to define...

Go's net/url.JoinPath treats every input as a path segment, so passing "/auth/logout?all_devices=true" produces "/api/v1/auth/logout%3Fall_devices=true" with the question mark URL-encoded into the pat...

While refactoring authorization dependencies, app startup failed with `fastapi.exceptions.FastAPIError: Invalid args for response field` because dependency parameters used repository classes directly....

In a FastAPI RBAC refactor, route-level checks scaled better when role hierarchy was centralized (`user < tenant_admin < platform_admin`) and minimum-role checks were generated from one dependency fac...

To test a local edit to a Claude Code plugin without cutting a release, copying into only `~/.claude/plugins/marketplaces/<name>/...` is insufficient; Claude Code also reads from `~/.claude/plugins/ca...

Playwright's `--update-snapshots` option accepts an optional mode value. In wrapper-script usage, a command like `... --update-snapshots e2e/spec.ts` can be parsed as mode=`e2e/spec.ts` and fail with ...

Per RFC 9110, the 422 reason phrase was renamed from Unprocessable Entity to Unprocessable Content; wire status remains 422. Starlette (verified 1.0.0, 2026-05) keeps the old constant as a deprecated ...

FastAPI's BackgroundTasks fire AFTER the response is sent and AFTER request-scoped dependencies with `yield` have run their cleanup, so a Session opened by a `with Session(engine) as session:` generat...

When uploading files via `fetch` with a `FormData` body, setting `Content-Type: multipart/form-data` manually (or via a Headers object that inherits a default JSON Content-Type) drops the auto-generat...

In FastAPI, attaching `response.set_cookie(...)` or `response.delete_cookie(...)` to the injected `Response` dependency *only* takes effect when the handler returns normally. When the handler raises H...

Most recent npm/PyPI compromises were detected and yanked within hours: debug+chalk (~2h live, 2025-09), litellm (~40min, 2026-03), axios (~3h, 2026-04), lottie-player (~2h, 2024-10), ua-parser-js (~4...

When a JSON Schema document uses $ref to point at another schema by its $id, the jsonschema Draft202012Validator (verified jsonschema 4.26.0, May 2026) will attempt to retrieve the referenced schema v...

Slack's `openid.connect.userInfo` endpoint returns Slack-specific claims (e.g. `https://slack.com/user_id`, `https://slack.com/team_id`, `https://slack.com/team_name`, `https://slack.com/team_domain`)...

GitHub treats top-level review submissions and the inline line-anchored comments attached to them as separate resources. `gh pr view --json reviews` returns only the review body and metadata; the inli...

Commonly used GitHub Actions advance major versions faster than LLM training data. Acting on LLM-suggested versions without verification may reference an outdated major that misses features, performan...

`git rebase -i --autosquash` normally opens `$GIT_EDITOR` even when every line in the todo list is `pick` or `fixup`, blocking automation. Setting `GIT_SEQUENCE_EDITOR=true` (or `:`) makes git use the...

pnpm can abort with ERR_PNPM_ABORTED_REMOVE_MODULES_DIR_NO_TTY during dependency checks when it needs to remove node_modules but no TTY is available. Verified on macOS with pnpm 11.1.1 during make-dri...

Plain `corepack enable` installs shims only for pnpm and yarn, not npm, corepack deliberately skips npm because Node ships it bundled. Without the npm shim, bare `npm` calls resolve to the bundled glo...

When a .env.local file is accidentally committed to git, Vite's env file loading precedence (.env.local > .env) causes it to silently override the .env file written by CI workflows. This results in lo...

`corepack install` without arguments reads the `packageManager` field from the current directory's `package.json` and installs that exact version into the corepack cache, activating it for subsequent ...

Lockfiles (package-lock.json, pnpm-lock.yaml, uv.lock) record the hash of whatever version `install` first saw. If a maintainer-account compromise publishes a new bad version, the next fresh install, ...

When pnpm generates allowBuilds entries in pnpm-workspace.yaml (via `pnpm approve-builds` or scaffolding), it may produce placeholder text like `set this to true or false` instead of actual boolean va...

Action tags are git tags and can be silently retargeted. A compromised maintainer can repoint @v4 at a malicious commit and every consumer pulls it on the next run. GitHub added an org/enterprise poli...

Unlike `pull_request`, `pull_request_target` runs against the base branch with the target repository's secrets available, while PR-author-controlled fields (title, body, branch name, head ref) feed in...

In a monorepo with multiple release workflows, a release.types:[published] trigger fires for every release regardless of tag prefix. Using push.tags with a glob filter avoids this but prevents workflo...

Running nektos/act on Apple M-series chips with --container-architecture linux/amd64 produces warnings and unreliable results. The default act images lack many tools that GitHub-hosted runners provide...

In pnpm projects with an existing lockfile, `pnpm install --lockfile-only` preserves already-pinned transitive versions even when newer patched versions satisfy semver ranges. To remediate advisories ...

`GET https://api.hubapi.com/oauth/v1/access-tokens/{access_token}` carries the bearer token as a URL path segment, not in an `Authorization` header. Standard HTTP error handling (e.g. `httpx.HTTPStatu...

pnpm moved per-workspace supply-chain controls (minimumReleaseAge, minimumReleaseAgeExclude, onlyBuiltDependencies) out of .npmrc into pnpm-workspace.yaml and the global config.yaml. Settings placed i...

When SQLAlchemy/SQLModel creates a table with an sa.Enum(...) column against PostgreSQL, the database auto-creates a backing named type (e.g. `passwordresetstatus`). op.drop_table removes the table bu...

In a FastAPI + Alembic project using Docker Compose Postgres, `alembic upgrade head` failed with `Can't locate revision identified by '<id>'` because the local `alembic_version` table referenced an ol...

Adding a new audit-snapshot column (one whose value should record what was true at link/observation time, e.g. an OAuth identity's verified tenant domain, an order's currency at purchase) to a table t...

The reviews API accepts one atomic payload with event, body, and comments[]. Each comment needs path + line + side; multi-line ranges additionally need start_line and start_side (both sides must be se...

apron-auth's `IdentityProfile` exposes `tenancies: tuple[TenancyContext, ...]` rather than per-provider claim fields. Each `TenancyContext` carries `id`, `name`, `domain`, `raw`, and a normalized `own...

The Microsoft Graph implementation of OIDC userinfo (`https://graph.microsoft.com/oidc/userinfo`) returns only `sub`, `name`, `family_name`, `given_name`, `picture`, `email`, it does not include `tid`...

Verified against Microsoft identity platform UserInfo docs (fetched 2026-05): UserInfo is hosted on Microsoft Graph (`GET/POST https://graph.microsoft.com/oidc/userinfo`) and Microsoft explicitly note...

The Notion OAuth token-grant response (returned during code exchange at `https://api.notion.com/v1/oauth/token`) includes `workspace_id`, `workspace_name`, AND `workspace_icon`. The runtime `GET /v1/u...

In a multi-provider OAuth service, all provider clients shared one state store. Exchanging the callback code with an arbitrary client was incorrect because token endpoints/auth methods differ per prov...

The Salesforce userinfo response (`<instance_url>/services/oauth2/userinfo`) returns a top-level `id` field of the form `https://MYDOMAINNAME.my.salesforce.com/id/<orgId>/<userId>`. The MyDomain host ...

While implementing OAuth state hardening in Python/boto3 (verified 2026-04), conditional updates failed with `ValidationException` because `ttl` and `consumed` were treated as reserved keywords in exp...

Verified against Notion API docs for `POST /v1/oauth/revoke` and Authorization guide (fetched 2026-05). Docs clearly define token revocation mechanics but do not explicitly state that revocation remov...

Verified against Slack Developer Docs for `auth.revoke` (fetched 2026-05): the response section states revoking a bot token does not uninstall the bot user or app; org-wide apps require admin removal ...

`GET https://api.atlassian.com/oauth/token/accessible-resources` returns an array of `{id, name, url, scopes, avatarUrl}` objects, one OAuth token can be valid for multiple Cloud sites simultaneously....

HeroUI React v3's Alert (from @heroui/react) takes a status prop, not color. Accepted values: default, accent, danger, success, warning. Content composes via Alert.Indicator, Alert.Content, Alert.Titl...

`cq` from `mozilla-ai/tap` is currently published as a cask (`Casks/cq.rb`) and not as a formula. Adding `mozilla-ai/tap/cq` under formulas causes drift and missing-formula errors in checks. Verified ...

Homebrew can fail `brew bundle` before package install begins when a cask DSL change makes an existing cask stanza invalid. In this case, `1password-cli` failed with: `generate_completions_from_execut...

The any-llm batch completion architecture requires every provider to convert native batch job representations to OpenAI's Batch type (from openai.types.Batch) and results to BatchResult (custom datacl...

SQLite databases running in WAL journal mode (default in modern SQLite, and what most application embeddings use) hold recent writes in `<dbname>-wal` and `<dbname>-shm` sidecar files until a checkpoi...

Two approaches exist for replying to a PR review comment. The dedicated endpoint `POST /repos/{owner}/{repo}/pulls/{pull_number}/comments/{comment_id}/replies` is the simplest: only `body` is required...

AWS Bedrock has two distinct service clients: 'bedrock-runtime' (for inference: converse, invoke_model) and 'bedrock' (for control-plane: create_model_invocation_job, list_foundation_models, etc.). Th...

When iterating over directories or files whose names begin with `-` (common with hash-prefixed cache paths, project-slug directories, or any auto-generated naming scheme that uses `-` as a separator a...

When implementing multi-step API flows (e.g. upload then set permissions then insert), cleanup of earlier steps on later failure must cover three distinct failure modes: non-2xx HTTP responses (checke...

In a Tailwind v4 project, Biome's CSS parser treats directives like `@custom-variant pointer-fine (@media (pointer: fine))` and `@theme` as syntax errors with the message "Tailwind-specific syntax is ...

When migrating a React+TypeScript frontend from ESLint (js.recommended + tseslint.recommended + react-hooks.recommended) to Biome (linter.rules.recommended: true), expect a wave of a11y errors that th...

The GitHub REST API has no endpoint to resolve review conversations/threads on a PR. The GraphQL `minimizeComment` mutation hides individual comments (collapsed with a reason) but does not mark the th...

The commonly used https://drive.google.com/uc?id={file_id} URL pattern can return an HTML interstitial page instead of the raw file, particularly in enterprise/Workspace environments with external sha...

`PUT /me/drive/root:/{filename}:/content` always creates or overwrites a file at the drive root, regardless of where the original file lives. For download-modify-upload workflows (e.g. editing a Power...

This file is the canonical shape of a commons-approved KU on disk: a propose.json-compatible payload, kebab-case slug filename, placed under approved/synthetic because Mozilla.ai has signed it off and...

Using `gh issue comment --body "...` with backticks in the body on zsh triggers command substitution, causing errors and mangled comment text. Verified with GitHub CLI (zsh) on 2026-05. Passing plain ...

While triaging GitHub issues via `gh`, `gh issue close <n> --reason not_planned` fails with an invalid argument error. The CLI expects exact reason strings: `completed`, `not planned`, or `duplicate`....

The Linear GraphQL `Organization.urlKey` field, used for the workspace URL slug (`linear.app/<urlKey>`), can be renamed by users at any time. Linear keeps the previous three urlKeys as redirects, so a...

A Vite plugin that uses `this.emitFile()` inside `buildStart` to produce a static asset (e.g. /.well-known/<file>) will trip the warning `context method emitFile() is not supported in serve mode. This...