vexor 0.21.1__py3-none-any.whl → 0.22.0__py3-none-any.whl

vexor/cache.py

@@ -5,9 +5,13 @@ from __future__ import annotations
 import hashlib
 import os
 import sqlite3
+from collections import OrderedDict
 from dataclasses import dataclass
+from contextlib import contextmanager
+from contextvars import ContextVar
 from datetime import datetime, timezone, timedelta
 from pathlib import Path
+from threading import Lock
 from typing import Iterable, Mapping, Sequence
 
 import numpy as np
@@ -16,10 +20,18 @@ from .utils import collect_files
 
 DEFAULT_CACHE_DIR = Path(os.path.expanduser("~")) / ".vexor"
 CACHE_DIR = DEFAULT_CACHE_DIR
+_CACHE_DIR_OVERRIDE: ContextVar[Path | None] = ContextVar(
+    "vexor_cache_dir_override",
+    default=None,
+)
 CACHE_VERSION = 6
 DB_FILENAME = "index.db"
 EMBED_CACHE_TTL_DAYS = 30
 EMBED_CACHE_MAX_ENTRIES = 50_000
+EMBED_MEMORY_CACHE_MAX_ENTRIES = 2_048
+
+_EMBED_MEMORY_CACHE: "OrderedDict[tuple[str, str], np.ndarray]" = OrderedDict()
+_EMBED_MEMORY_LOCK = Lock()
 
 
 @dataclass(slots=True)
@@ -84,6 +96,55 @@ def embedding_cache_key(text: str) -> str:
     return hashlib.sha1(clean_text.encode("utf-8")).hexdigest()
 
 
+def _clear_embedding_memory_cache() -> None:
+    if EMBED_MEMORY_CACHE_MAX_ENTRIES <= 0:
+        return
+    with _EMBED_MEMORY_LOCK:
+        _EMBED_MEMORY_CACHE.clear()
+
+
+def _load_embedding_memory_cache(
+    model: str,
+    text_hashes: Sequence[str],
+) -> dict[str, np.ndarray]:
+    if EMBED_MEMORY_CACHE_MAX_ENTRIES <= 0:
+        return {}
+    results: dict[str, np.ndarray] = {}
+    with _EMBED_MEMORY_LOCK:
+        for text_hash in text_hashes:
+            if not text_hash:
+                continue
+            key = (model, text_hash)
+            vector = _EMBED_MEMORY_CACHE.pop(key, None)
+            if vector is None:
+                continue
+            _EMBED_MEMORY_CACHE[key] = vector
+            results[text_hash] = vector
+    return results
+
+
+def _store_embedding_memory_cache(
+    *,
+    model: str,
+    embeddings: Mapping[str, np.ndarray],
+) -> None:
+    if EMBED_MEMORY_CACHE_MAX_ENTRIES <= 0 or not embeddings:
+        return
+    with _EMBED_MEMORY_LOCK:
+        for text_hash, vector in embeddings.items():
+            if not text_hash:
+                continue
+            array = np.asarray(vector, dtype=np.float32)
+            if array.size == 0:
+                continue
+            key = (model, text_hash)
+            if key in _EMBED_MEMORY_CACHE:
+                _EMBED_MEMORY_CACHE.pop(key, None)
+            _EMBED_MEMORY_CACHE[key] = array
+        while len(_EMBED_MEMORY_CACHE) > EMBED_MEMORY_CACHE_MAX_ENTRIES:
+            _EMBED_MEMORY_CACHE.popitem(last=False)
+
+
 def _serialize_extensions(extensions: Sequence[str] | None) -> str:
     if not extensions:
         return ""
@@ -115,9 +176,32 @@ def _chunk_values(values: Sequence[object], size: int) -> Iterable[Sequence[obje
         yield values[idx : idx + size]
 
 
+def _resolve_cache_dir() -> Path:
+    override = _CACHE_DIR_OVERRIDE.get()
+    return override if override is not None else CACHE_DIR
+
+
+@contextmanager
+def cache_dir_context(path: Path | str | None):
+    """Temporarily override the cache directory for the current context."""
+
+    if path is None:
+        yield
+        return
+    dir_path = Path(path).expanduser().resolve()
+    if dir_path.exists() and not dir_path.is_dir():
+        raise NotADirectoryError(f"Path is not a directory: {dir_path}")
+    token = _CACHE_DIR_OVERRIDE.set(dir_path)
+    try:
+        yield
+    finally:
+        _CACHE_DIR_OVERRIDE.reset(token)
+
+
 def ensure_cache_dir() -> Path:
-    CACHE_DIR.mkdir(parents=True, exist_ok=True)
-    return CACHE_DIR
+    cache_dir = _resolve_cache_dir()
+    cache_dir.mkdir(parents=True, exist_ok=True)
+    return cache_dir
 
 
 def set_cache_dir(path: Path | str | None) -> None:
@@ -134,8 +218,8 @@ def set_cache_dir(path: Path | str | None) -> None:
 def cache_db_path() -> Path:
     """Return the absolute path to the shared SQLite cache database."""
 
-    ensure_cache_dir()
-    return CACHE_DIR / DB_FILENAME
+    cache_dir = ensure_cache_dir()
+    return cache_dir / DB_FILENAME
 
 
 def cache_file(root: Path, model: str, include_hidden: bool) -> Path:  # pragma: no cover - kept for API parity
@@ -1310,19 +1394,23 @@ def load_embedding_cache(
     unique_hashes = list(dict.fromkeys([value for value in text_hashes if value]))
     if not unique_hashes:
         return {}
+    results = _load_embedding_memory_cache(model, unique_hashes)
+    missing = [value for value in unique_hashes if value not in results]
+    if not missing:
+        return results
     db_path = cache_db_path()
     owns_connection = conn is None
     try:
         connection = conn or _connect(db_path, readonly=True)
     except sqlite3.OperationalError:
-        return {}
+        return results
     try:
         try:
             _ensure_schema_readonly(connection, tables=("embedding_cache",))
         except sqlite3.OperationalError:
-            return {}
-        results: dict[str, np.ndarray] = {}
-        for chunk in _chunk_values(unique_hashes, 900):
+            return results
+        disk_results: dict[str, np.ndarray] = {}
+        for chunk in _chunk_values(missing, 900):
             placeholders = ", ".join("?" for _ in chunk)
             rows = connection.execute(
                 f"""
@@ -1339,7 +1427,10 @@ def load_embedding_cache(
                 vector = np.frombuffer(blob, dtype=np.float32)
                 if vector.size == 0:
                     continue
-                results[row["text_hash"]] = vector
+                disk_results[row["text_hash"]] = vector
+        if disk_results:
+            _store_embedding_memory_cache(model=model, embeddings=disk_results)
+            results.update(disk_results)
         return results
     finally:
         if owns_connection:
@@ -1356,6 +1447,7 @@ def store_embedding_cache(
 
     if not embeddings:
         return
+    _store_embedding_memory_cache(model=model, embeddings=embeddings)
     db_path = cache_db_path()
     owns_connection = conn is None
     connection = conn or _connect(db_path)

vexor/config.py

@@ -5,6 +5,8 @@ from __future__ import annotations
 import json
 import os
 from dataclasses import dataclass
+from contextlib import contextmanager
+from contextvars import ContextVar
 from collections.abc import Mapping
 from pathlib import Path
 from typing import Any, Dict
@@ -15,6 +17,10 @@ from .text import Messages
 DEFAULT_CONFIG_DIR = Path(os.path.expanduser("~")) / ".vexor"
 CONFIG_DIR = DEFAULT_CONFIG_DIR
 CONFIG_FILE = CONFIG_DIR / "config.json"
+_CONFIG_DIR_OVERRIDE: ContextVar[Path | None] = ContextVar(
+    "vexor_config_dir_override",
+    default=None,
+)
 DEFAULT_MODEL = "text-embedding-3-small"
 DEFAULT_GEMINI_MODEL = "gemini-embedding-001"
 DEFAULT_LOCAL_MODEL = "intfloat/multilingual-e5-small"
@@ -74,10 +80,40 @@ def _parse_remote_rerank(raw: object) -> RemoteRerankConfig | None:
     )
 
 
+def _resolve_config_dir() -> Path:
+    override = _CONFIG_DIR_OVERRIDE.get()
+    return override if override is not None else CONFIG_DIR
+
+
+def _resolve_config_file() -> Path:
+    override = _CONFIG_DIR_OVERRIDE.get()
+    if override is not None:
+        return override / "config.json"
+    return CONFIG_FILE
+
+
+@contextmanager
+def config_dir_context(path: Path | str | None):
+    """Temporarily override the config directory for the current context."""
+
+    if path is None:
+        yield
+        return
+    dir_path = Path(path).expanduser().resolve()
+    if dir_path.exists() and not dir_path.is_dir():
+        raise NotADirectoryError(f"Path is not a directory: {dir_path}")
+    token = _CONFIG_DIR_OVERRIDE.set(dir_path)
+    try:
+        yield
+    finally:
+        _CONFIG_DIR_OVERRIDE.reset(token)
+
+
 def load_config() -> Config:
-    if not CONFIG_FILE.exists():
+    config_file = _resolve_config_file()
+    if not config_file.exists():
         return Config()
-    raw = json.loads(CONFIG_FILE.read_text(encoding="utf-8"))
+    raw = json.loads(config_file.read_text(encoding="utf-8"))
     rerank = (raw.get("rerank") or DEFAULT_RERANK).strip().lower()
     if rerank not in SUPPORTED_RERANKERS:
         rerank = DEFAULT_RERANK
@@ -101,7 +137,8 @@ def load_config() -> Config:
 
 
 def save_config(config: Config) -> None:
-    CONFIG_DIR.mkdir(parents=True, exist_ok=True)
+    config_dir = _resolve_config_dir()
+    config_dir.mkdir(parents=True, exist_ok=True)
     data: Dict[str, Any] = {}
     if config.api_key:
         data["api_key"] = config.api_key
@@ -130,15 +167,19 @@ def save_config(config: Config) -> None:
             remote_data["model"] = config.remote_rerank.model
         if remote_data:
             data["remote_rerank"] = remote_data
-    CONFIG_FILE.write_text(json.dumps(data, ensure_ascii=False, indent=2), encoding="utf-8")
+    config_file = _resolve_config_file()
+    config_file.write_text(
+        json.dumps(data, ensure_ascii=False, indent=2),
+        encoding="utf-8",
+    )
 
 
 def local_model_dir() -> Path:
-    return CONFIG_DIR / "models"
+    return _resolve_config_dir() / "models"
 
 
 def flashrank_cache_dir(*, create: bool = True) -> Path:
-    cache_dir = CONFIG_DIR / "flashrank"
+    cache_dir = _resolve_config_dir() / "flashrank"
     if create:
         cache_dir.mkdir(parents=True, exist_ok=True)
     return cache_dir

vexor/services/content_extract_service.py

@@ -108,6 +108,12 @@ TEXT_EXTENSIONS = (
     ".vb",
     ".ps1",
     ".bash",
+    ".zsh",
+    ".fish",
+    ".vue",
+    ".jsx",
+    ".tsx",
+    ".scss",
 )
 
 

vexor/services/search_service.py

@@ -689,35 +689,22 @@ def perform_search(request: SearchRequest) -> SearchResponse:
     )
 
 
-def _perform_search_with_temporary_index(request: SearchRequest) -> SearchResponse:
-    from .index_service import build_index_in_memory  # local import
-
-    paths, file_vectors, metadata = build_index_in_memory(
-        request.directory,
-        include_hidden=request.include_hidden,
-        respect_gitignore=request.respect_gitignore,
-        mode=request.mode,
-        recursive=request.recursive,
-        model_name=request.model_name,
-        batch_size=request.batch_size,
-        embed_concurrency=request.embed_concurrency,
-        extract_concurrency=request.extract_concurrency,
-        extract_backend=request.extract_backend,
-        provider=request.provider,
-        base_url=request.base_url,
-        api_key=request.api_key,
-        local_cuda=request.local_cuda,
-        exclude_patterns=request.exclude_patterns,
-        extensions=request.extensions,
-        no_cache=request.no_cache,
-    )
+def search_from_vectors(
+    request: SearchRequest,
+    *,
+    paths: Sequence[Path],
+    file_vectors: np.ndarray,
+    metadata: dict,
+    is_stale: bool = False,
+) -> SearchResponse:
+    """Return ranked results from an in-memory index."""
 
     if not len(paths):
         return SearchResponse(
             base_path=request.directory,
             backend=None,
             results=[],
-            is_stale=False,
+            is_stale=is_stale,
             index_empty=True,
         )
 
@@ -813,12 +800,43 @@ def _perform_search_with_temporary_index(request: SearchRequest) -> SearchRespon
         base_path=request.directory,
         backend=searcher.device,
         results=results,
-        is_stale=False,
+        is_stale=is_stale,
         index_empty=False,
         reranker=reranker,
     )
 
 
+def _perform_search_with_temporary_index(request: SearchRequest) -> SearchResponse:
+    from .index_service import build_index_in_memory  # local import
+
+    paths, file_vectors, metadata = build_index_in_memory(
+        request.directory,
+        include_hidden=request.include_hidden,
+        respect_gitignore=request.respect_gitignore,
+        mode=request.mode,
+        recursive=request.recursive,
+        model_name=request.model_name,
+        batch_size=request.batch_size,
+        embed_concurrency=request.embed_concurrency,
+        extract_concurrency=request.extract_concurrency,
+        extract_backend=request.extract_backend,
+        provider=request.provider,
+        base_url=request.base_url,
+        api_key=request.api_key,
+        local_cuda=request.local_cuda,
+        exclude_patterns=request.exclude_patterns,
+        extensions=request.extensions,
+        no_cache=request.no_cache,
+    )
+    return search_from_vectors(
+        request,
+        paths=paths,
+        file_vectors=file_vectors,
+        metadata=metadata,
+        is_stale=False,
+    )
+
+
 def _load_index_vectors_for_request(
     request: SearchRequest,
     *,

{vexor-0.21.1.dist-info → vexor-0.22.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: vexor
-Version: 0.21.1
+Version: 0.22.0
 Summary: A vector-powered CLI for semantic search over files.
 Project-URL: Repository, https://github.com/scarletkc/vexor
 Author: scarletkc
@@ -76,6 +76,13 @@ It supports configurable embedding and reranking providers, and exposes the same
       Vexor Demo Video
     </video>
 
+## Featured In
+
+Vexor has been recognized and featured by the community:
+
+- **[Ruan Yifeng's Weekly (Issue #379)](https://github.com/ruanyf/weekly/blob/master/docs/issue-379.md#ai-%E7%9B%B8%E5%85%B3)** - A leading tech newsletter in the Chinese developer community.
+- **[Awesome Claude Skills](https://github.com/VoltAgent/awesome-claude-skills?tab=readme-ov-file#development-and-testing)** - Curated list of best-in-class skills for AI agents.
+
 ## Why Vexor?
 
 When you remember what a file *does* but forget its name or location, Vexor finds it instantly—no grep patterns or directory traversal needed.
@@ -315,8 +322,14 @@ Porcelain output fields: `rank`, `similarity`, `path`, `chunk_index`, `start_lin
 
 See [docs](https://github.com/scarletkc/vexor/tree/main/docs) for more details.
 
+## Contributing
+
 Contributions, issues, and PRs welcome! Star if you find it helpful.
 
+## Star History
+
+[![Star History Chart](https://api.star-history.com/svg?repos=scarletkc/vexor&type=date&legend=top-left)](https://www.star-history.com/#scarletkc/vexor&type=date&legend=top-left)
+
 ## License
 
 [MIT](http://github.com/scarletkc/vexor/blob/main/LICENSE)

{vexor-0.21.1.dist-info → vexor-0.22.0.dist-info}/RECORD

@@ -1,9 +1,9 @@
-vexor/__init__.py,sha256=Ab63nROf2nbDW-xY4wuNU_DS0K8hsqfPa1KjvCaKJzA,441
+vexor/__init__.py,sha256=EQXPbwsXfHJAK3mNlhYfc4UwVqOMqL52cO46xg1GNCo,632
 vexor/__main__.py,sha256=ZFzom1wCfP6TPXe3aoDFpNcUgjbCZ7Quy_vfzNsH5Fw,426
-vexor/api.py,sha256=YCHpiydbPbRJUqdQYrpwe1JrRI-w_7LRuyZDGBP1_d4,11506
-vexor/cache.py,sha256=20SaiBKkPJIDXHtflX6uHiQXI4DtD6wx7RtWbz2l6LU,54339
+vexor/api.py,sha256=W6eJLNbg5uBC2gcE8kq34iUciJCYZjY9Zsytxunl-vs,35860
+vexor/cache.py,sha256=irCGy5XIcRKX5EFk7plKDVqGHrTjRpaWgXnFDiVChXk,57323
 vexor/cli.py,sha256=M9GKdD_mJ068Zpm62znTp0KhhKp1dkh_WHmfJHR9hwU,68094
-vexor/config.py,sha256=CiPfEH7Ilt6XepEx4p02qfW5HfkpNDBjhEMyckbSWaA,17413
+vexor/config.py,sha256=CEL5u7afZV81Y0i9FVsj8GVzZlb2C2gC17r5lNKu-aM,18570
 vexor/modes.py,sha256=N_wAWoqbxmCfko-v520p59tpAYvUwraCSSQRtMaF4ac,11549
 vexor/output.py,sha256=iooZgLlK8dh7ajJ4XMHUNNx0qyTVtD_OAAwrBx5MeqE,864
 vexor/search.py,sha256=MSU4RmH6waFYOofkIdo8_ElTiz1oNaKuvr-3umif7Bs,6826
@@ -16,18 +16,18 @@ vexor/providers/openai.py,sha256=YnJDY9gJW7RfGGdkgswVHvmOKNvgLRQUsbpA1MUuLPg,535
 vexor/services/__init__.py,sha256=dA_i2N03vlYmbZbEK2knzJLWviunkNWbzN2LWPNvMk0,160
 vexor/services/cache_service.py,sha256=ywt6AgupCJ7_wC3je4znCMw5_VBouw3skbDTAt8xw6o,1639
 vexor/services/config_service.py,sha256=PojolfbSKh9pW8slF4qxCOs9hz5L6xvjf_nB7vfVlsU,5039
-vexor/services/content_extract_service.py,sha256=zdhLxpNv70BU7irLf3Uc0ou9rKSvdjtrDcHkgRKlMn4,26421
+vexor/services/content_extract_service.py,sha256=oO7Hbadwp3uiyqCbr_4MRXQsUeMix2D98i-Yp94PwFk,26495
 vexor/services/index_service.py,sha256=FXf1bBoqj4-K1l38ItxHf6Oh7QHVIdNAdVY2kg_Zoq8,32265
 vexor/services/init_service.py,sha256=3D04hylGA9FRQhLHCfR95nMko3vb5MNBcRb9nWWaUE8,26863
 vexor/services/js_parser.py,sha256=eRtW6KlK4JBYDGbyoecHVqLZ0hcx-Cc0kx6bOujHPAQ,16254
 vexor/services/keyword_service.py,sha256=vmke8tII9kTwRDdBaLHBc6Hpy_B3p98L65iGkCQgtMU,2211
-vexor/services/search_service.py,sha256=K7SiAuMA7bGeyPWOHPMKpFFvzzkj5kHWwa3p94NakJs,38663
+vexor/services/search_service.py,sha256=-7qHfebMOmXWiVBVYoBji6eaZr8OOY3q1vbtJaY3I6E,39076
 vexor/services/skill_service.py,sha256=Rrgt3OMsKPPiXOiRhSNAWjBM9UNz9qmSWQe3uYGzq4M,4863
 vexor/services/system_service.py,sha256=KPlv83v3rTvBiNiH7vrp6tDmt_AqHxuUd-5RI0TfvWs,24638
 vexor/_bundled_skills/vexor-cli/SKILL.md,sha256=m3FlyqgHBdRwyGPEp8PrUS21K0G2jEl88tRvhSPta08,2798
 vexor/_bundled_skills/vexor-cli/references/install-vexor.md,sha256=IUBShLI1mAxugwUIMAJQ5_j6KcaPWfobe0gSd6MWU7w,1245
-vexor-0.21.1.dist-info/METADATA,sha256=jS_xdqPXD8WsDNKd684w5eHmj_f1CHvNMR-DY-MvBQg,13494
-vexor-0.21.1.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
-vexor-0.21.1.dist-info/entry_points.txt,sha256=dvxp6Q1R1d6bozR7TwmpdJ0X_v83MkzsLPagGY_lfr0,40
-vexor-0.21.1.dist-info/licenses/LICENSE,sha256=wP7TAKRll1t9LoYGxWS9NikPM_0hCc00LmlLyvQBsL8,1066
-vexor-0.21.1.dist-info/RECORD,,
+vexor-0.22.0.dist-info/METADATA,sha256=UAqD6ciQCaP3eBrOGA3unO_XLX0eArcYHfcXuRhjN8c,14154
+vexor-0.22.0.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
+vexor-0.22.0.dist-info/entry_points.txt,sha256=dvxp6Q1R1d6bozR7TwmpdJ0X_v83MkzsLPagGY_lfr0,40
+vexor-0.22.0.dist-info/licenses/LICENSE,sha256=wP7TAKRll1t9LoYGxWS9NikPM_0hCc00LmlLyvQBsL8,1066
+vexor-0.22.0.dist-info/RECORD,,