hecvec 0.4.3__tar.gz → 0.4.4__tar.gz

{hecvec-0.4.3 → hecvec-0.4.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: hecvec
-Version: 0.4.3
+Version: 0.4.4
 Summary: List directories (safe root), filter .txt/.md files, read as text, chunk, embed, and push to Chroma.
 License-Expression: MIT
 Keywords: chunking,document-pipeline,listdir,text-files

{hecvec-0.4.3 → hecvec-0.4.4}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "hecvec"
-version = "0.4.3"
+version = "0.4.4"
 description = "List directories (safe root), filter .txt/.md files, read as text, chunk, embed, and push to Chroma."
 readme = "README.md"
 requires-python = ">=3.9,<3.14"

{hecvec-0.4.3 → hecvec-0.4.4}/src/hecvec/__init__.py

@@ -2,7 +2,7 @@
 hecvec: modular library — listdir, read, chunk (token/recursive), embed, Chroma. No API.
 """
 from hecvec.chunking import chunk_documents, chunk_text
-from hecvec.chroma_client import add_documents, get_client, get_or_create_collection
+from hecvec.chroma_client import add_documents, get_client, get_or_create_collection, ChromaMode
 from hecvec.chunkers import ChunkingMethod, chunk_documents as chunk_documents_by_method
 from hecvec.chroma_list import list_collections
 from hecvec.embeddings import embed_texts
@@ -21,6 +21,7 @@ __all__ = [
     "ReadText",
     "add_documents",
     "chunk_documents",
+    "ChromaMode",
     "chunk_documents_by_method",
     "chunk_text",
     "embed_texts",
@@ -34,4 +35,4 @@ __all__ = [
     "__version__",
 ]
 
-__version__ = "0.4.3"
+__version__ = "0.4.4"

{hecvec-0.4.3 → hecvec-0.4.4}/src/hecvec/chroma_client.py

@@ -4,7 +4,7 @@ Requires: pip install hecvec[chroma] (chromadb).
 """
 from __future__ import annotations
 
-from typing import TYPE_CHECKING
+from typing import TYPE_CHECKING, Literal
 
 if TYPE_CHECKING:
     import chromadb
@@ -12,24 +12,33 @@ if TYPE_CHECKING:
 DEFAULT_HOST = "localhost"
 DEFAULT_PORT = 8000
 
+ChromaMode = Literal["auto", "server", "ephemeral"]
+
 
 def get_client(
     host: str = DEFAULT_HOST,
     port: int = DEFAULT_PORT,
     persist_path: str | None = None,
+    mode: ChromaMode = "auto",
 ):
     """
-    Return a Chroma client and mode.
+    Return a Chroma client and mode. All backends use the same Chroma API (same add/query methods).
 
-    - If persist_path is set: use PersistentClient(path). Data is stored on disk.
-      Returns (client, "persistent"). host/port are ignored.
-    - Else: try HttpClient(host, port). If that fails, use EphemeralClient().
-      Returns (client, "server") or (client, "ephemeral").
+    - persist_path set: use PersistentClient(path). Data on disk. Returns (client, "persistent").
+    - mode "ephemeral": use EphemeralClient(). In-memory, no persistence. Returns (client, "ephemeral").
+    - mode "server": use HttpClient(host, port) only. Raises if server unreachable. Returns (client, "server").
+    - mode "auto": try server, then fall back to ephemeral. Returns (client, "server") or (client, "ephemeral").
     """
     import chromadb
     if persist_path is not None:
         client = chromadb.PersistentClient(path=persist_path)
         return client, "persistent"
+    if mode == "ephemeral":
+        return chromadb.EphemeralClient(), "ephemeral"
+    if mode == "server":
+        client = chromadb.HttpClient(host=host, port=port)
+        return client, "server"
+    # auto: try server, fall back to ephemeral
     try:
         client = chromadb.HttpClient(host=host, port=port)
         return client, "server"

{hecvec-0.4.3 → hecvec-0.4.4}/src/hecvec/pipeline.py

@@ -11,6 +11,7 @@ from time import perf_counter
 from typing import Any
 
 from hecvec.chroma_client import add_documents, get_client
+from hecvec.chroma_client import ChromaMode
 from hecvec.chunkers import ChunkingMethod, chunk_documents as chunk_documents_by_method
 from hecvec.embeddings import embed_texts
 from hecvec.env import load_openai_key
@@ -68,6 +69,8 @@ class Slicer:
         self.collection_name = collection_name
         self.chroma_host = chroma_host
         self.chroma_port = chroma_port
+        self.chroma_mode = chroma_mode
+        self.chroma_persist_path = chroma_persist_path
         self.embedding_model = embedding_model
         self.chunk_size = chunk_size
         self.chunk_overlap = chunk_overlap
@@ -88,6 +91,8 @@ class Slicer:
             collection_name=kwargs.pop("collection_name", self.collection_name),
             chroma_host=kwargs.pop("chroma_host", self.chroma_host),
             chroma_port=kwargs.pop("chroma_port", self.chroma_port),
+            chroma_mode=kwargs.pop("chroma_mode", self.chroma_mode),
+            chroma_persist_path=kwargs.pop("chroma_persist_path", self.chroma_persist_path),
             embedding_model=kwargs.pop("embedding_model", self.embedding_model),
             chunk_size=kwargs.pop("chunk_size", self.chunk_size),
             chunk_overlap=kwargs.pop("chunk_overlap", self.chunk_overlap),
@@ -108,6 +113,8 @@ class Slicer:
         collection_name: str = "hecvec",
         chroma_host: str = "localhost",
         chroma_port: int = 8000,
+        chroma_mode: ChromaMode = "auto",
+        chroma_persist_path: str | Path | None = None,
         embedding_model: str = "text-embedding-3-small",
         chunk_size: int = 200,
         chunk_overlap: int = 0,
@@ -209,11 +216,18 @@ class Slicer:
         # 5/5 Push to Chroma
         stage_start = perf_counter()
         logger.info("[5/5] Writing to Chroma | host=%s | port=%s | collection=%s", chroma_host, chroma_port, collection_name)
-        client, chroma_mode = get_client(host=chroma_host, port=chroma_port)
-        if chroma_mode == "server":
+        client, chroma_mode_used = get_client(
+            host=chroma_host,
+            port=chroma_port,
+            persist_path=str(chroma_persist_path) if chroma_persist_path is not None else None,
+            mode=chroma_mode,
+        )
+        if chroma_mode_used == "server":
             logger.info("[5/5] Using Chroma server at %s:%s", chroma_host, chroma_port)
+        elif chroma_mode_used == "persistent":
+            logger.info("[5/5] Using Chroma persistent storage at %s", chroma_persist_path)
         else:
-            logger.info("[5/5] Chroma server unreachable; using in-memory (ephemeral) client")
+            logger.info("[5/5] Using Chroma in-memory (ephemeral) client")
         logger.info("[5/5] Adding %d document chunk(s) to collection...", len(documents))
         add_result = add_documents(client, collection_name, ids, embeddings, documents)
         if add_result["collection_existed"]: