Fix by linteres

Author: Ariel Vernaza

.github/workflows/tests.yml

@@ -7,6 +7,29 @@ on:
     branches: [main, master]
 
 jobs:
+  lint:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+
+      - name: Install uv
+        uses: astral-sh/setup-uv@v7
+        with:
+          enable-cache: true
+
+      - name: Set up Python
+        run: uv python install
+
+      - name: Install dependencies
+        run: uv sync
+
+      - name: Ruff check
+        run: uv run ruff check .
+
+      - name: Ruff format check
+        run: uv run ruff format --check .
+
   test:
     runs-on: ubuntu-latest
     steps:

README.md

@@ -271,7 +271,20 @@ uv run pytest -v
 ```
 
 **4. CI (GitHub Actions)**  
-On every push and pull request to `main` or `master`, the workflow in `.github/workflows/tests.yml` runs `uv sync` and `uv run pytest -v`. No secrets required; embedding tests that need optional deps are skipped in CI.
+On every push and pull request to `main` or `master`, the workflow in `.github/workflows/tests.yml` runs **lint** (Ruff check + format check) and **tests** (`uv run pytest -v`). No secrets required; embedding tests that need optional deps are skipped in CI.
+
+**5. Linting (Ruff)**  
+Locally: `uv run ruff check .` and `uv run ruff format --check .`. To auto-fix: `uv run ruff check . --fix` and `uv run ruff format .`. Config is in `pyproject.toml` under `[tool.ruff]`.
+
+**6. Linting desde Docker Compose**  
+Construir la imagen de test (incluye Ruff) y ejecutar lint sobre el código montado:
+
+```bash
+docker compose --profile lint build lint
+docker compose --profile lint run --rm lint
+```
+
+(Con Podman: `podman compose --profile lint build lint` y `podman compose --profile lint run --rm lint`.) El servicio `lint` monta `app/`, `tests/` y `pyproject.toml`; no hace falta reconstruir para lintar cambios. Para ejecutar tests en el mismo contenedor: `docker compose --profile lint run --rm lint uv run pytest -v`.
 
 ---
 

app/api/schemas.py

@@ -36,7 +36,11 @@ class ChunkCreate(BaseModel):
     """
 
     text: str = Field(..., min_length=1, description="Chunk text content.")
-    embedding: list[float] = Field(..., min_length=1, description="Vector embedding (e.g. from Cohere or Sentence Transformers).")
+    embedding: list[float] = Field(
+        ...,
+        min_length=1,
+        description="Vector embedding (e.g. from Cohere or Sentence Transformers).",
+    )
     name: str | None = Field(default=None, description="Optional label for the chunk.")
 
 
@@ -60,7 +64,9 @@ class IndexRequest(BaseModel):
 class SearchRequest(BaseModel):
     """Request body for POST /libraries/{id}/search. Query vector and number of neighbors."""
 
-    embedding: list[float] = Field(..., min_length=1, description="Query vector (same dimension as indexed chunks).")
+    embedding: list[float] = Field(
+        ..., min_length=1, description="Query vector (same dimension as indexed chunks)."
+    )
     k: int = Field(..., ge=1, le=1000, description="Number of nearest neighbors to return.")
 
 
@@ -74,4 +80,6 @@ class SearchResultItem(BaseModel):
 class SearchResponse(BaseModel):
     """Response for POST /libraries/{id}/search. List of (chunk_id, distance) sorted by distance ascending."""
 
-    results: list[SearchResultItem] = Field(..., description="Up to k nearest chunks with their distances.")
+    results: list[SearchResultItem] = Field(
+        ..., description="Up to k nearest chunks with their distances."
+    )

app/core/concurrency.py

@@ -9,8 +9,8 @@
 from __future__ import annotations
 
 import threading
+from collections.abc import Generator
 from contextlib import contextmanager
-from typing import Generator
 
 
 class RWLock:

app/core/embedding/cohere.py

@@ -7,9 +7,10 @@
 
 from __future__ import annotations
 
-import httpx
 from typing import Any
 
+import httpx
+
 # Default Cohere model dimensions (embed-english-v3.0, embed-multilingual-v3.0)
 COHERE_EMBED_DIM = 1024
 
@@ -64,4 +65,4 @@ def _embed(self, texts: list[str], input_type: str) -> list[list[float]]:
             r.raise_for_status()
             data: dict[str, Any] = r.json()
         embeddings = data.get("embeddings", [])
-        return [e for e in embeddings]
+        return list(embeddings)

app/core/embedding/image.py

@@ -22,6 +22,7 @@ def _lazy_import() -> tuple[object, object, object, object]:
     import torch  # type: ignore[import-untyped]
     from PIL import Image  # type: ignore[import-untyped]
     from transformers import AutoModel, AutoProcessor  # type: ignore[import-untyped]
+
     return AutoModel, AutoProcessor, torch, Image
 
 
@@ -66,7 +67,9 @@ def dimension(self) -> int:
 
     def embed_documents(self, texts: list[str]) -> list[list[float]]:
         """Not supported: use a text embedder (Cohere or Transformer) for text chunks."""
-        raise NotImplementedError("ImageEmbedder only supports embed_images; use a text embedder for documents.")
+        raise NotImplementedError(
+            "ImageEmbedder only supports embed_images; use a text embedder for documents."
+        )
 
     def embed_queries(self, texts: list[str]) -> list[list[float]]:
         """Not supported: use embed_images for image query."""
@@ -76,8 +79,8 @@ def embed_images(self, images: list[bytes]) -> list[list[float]]:
         """Indexing/search: embed image chunks (raw bytes, e.g. from PDF image extraction). Each item is one image."""
         if not images:
             return []
-        from PIL import Image  # type: ignore[import-untyped]
         import torch  # type: ignore[import-untyped]
+        from PIL import Image  # type: ignore[import-untyped]
 
         pil_images = [Image.open(io.BytesIO(img)).convert("RGB") for img in images]
         processor, model, device = self._get_processor_and_model()

app/core/embedding/registry.py

@@ -18,8 +18,8 @@
 
 from app.core.embedding.base import Embedder, EmbedderKind
 from app.core.embedding.cohere import CohereEmbedder
-from app.core.embedding.transformer import TransformerEmbedder
 from app.core.embedding.image import ImageEmbedder
+from app.core.embedding.transformer import TransformerEmbedder
 
 
 def _parse_embedder_uri(uri: str) -> tuple[EmbedderKind, str | None]:
@@ -54,11 +54,15 @@ def get_embedder(
             raise ValueError("COHERE_API_KEY is required for cohere embedder")
         return CohereEmbedder(api_key=api_key)
     if kind == EmbedderKind.EMBEDDING_TRANSFORMER:
-        model_name = model or os.environ.get("EMBEDDING_TRANSFORMER_MODEL", "sentence-transformers/all-MiniLM-L6-v2")
+        model_name = model or os.environ.get(
+            "EMBEDDING_TRANSFORMER_MODEL", "sentence-transformers/all-MiniLM-L6-v2"
+        )
         dev = device or os.environ.get("EMBEDDING_DEVICE")
         return TransformerEmbedder(model_name=model_name, device=dev)
     if kind == EmbedderKind.EMBEDDING_IMAGE:
-        model_name = model or os.environ.get("EMBEDDING_IMAGE_MODEL", "openai/clip-vit-base-patch32")
+        model_name = model or os.environ.get(
+            "EMBEDDING_IMAGE_MODEL", "openai/clip-vit-base-patch32"
+        )
         dev = device or os.environ.get("EMBEDDING_DEVICE")
         return ImageEmbedder(model_name=model_name, device=dev)
     raise ValueError(f"Unknown embedder URI: {uri}")

app/core/embedding/transformer.py

@@ -20,6 +20,7 @@
 
 def _lazy_import() -> tuple[object, object]:
     import sentence_transformers  # type: ignore[import-untyped]
+
     return sentence_transformers.SentenceTransformer, sentence_transformers  # type: ignore[attr-defined,no-any-return]
 
 

app/indexes/ivf.py

@@ -93,7 +93,9 @@ def build(
         for i, label in enumerate(labels):
             cluster_ids[label].append(ids[i])
             cluster_vectors[label].append(vectors[i])
-        cluster_vectors_arr = [np.array(cv, dtype=np.float64) if cv else np.empty((0, d)) for cv in cluster_vectors]
+        cluster_vectors_arr = [
+            np.array(cv, dtype=np.float64) if cv else np.empty((0, d)) for cv in cluster_vectors
+        ]
         return cls(cluster_ids, cluster_vectors_arr, centroids, n_probe)
 
     def search(self, query: list[float], k: int) -> SearchResult:

app/indexes/kd_tree.py

@@ -19,13 +19,15 @@
 @dataclass
 class _Leaf:
     """Leaf node: holds a single point (id, vector)."""
+
     id: UUID
     vector: np.ndarray
 
 
 @dataclass
 class _Internal:
     """Internal node: split dimension and value, left and right subtrees."""
+
     split_dim: int
     split_value: float
     left: _Leaf | _Internal