chunksilo 2.3.3__tar.gz → 2.4.0__tar.gz

{chunksilo-2.3.3/src/chunksilo.egg-info → chunksilo-2.4.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: chunksilo
-Version: 2.3.3
+Version: 2.4.0
 Summary: Local RAG-based semantic document search with MCP server interface
 Author: Fredrik Reveny
 License-Expression: Apache-2.0
@@ -30,12 +30,13 @@ Requires-Dist: huggingface-hub<2,>=0.22.0
 Requires-Dist: flashrank<1,>=0.1.0
 Requires-Dist: fastembed<1,>=0.5.0
 Requires-Dist: pyyaml<7,>=6.0
+Requires-Dist: llama-index-readers-confluence<1,>=0.6.0
+Requires-Dist: jira<4,>=3.5.0
 Provides-Extra: confluence
-Requires-Dist: llama-index-readers-confluence<1,>=0.6.0; extra == "confluence"
 Provides-Extra: jira
-Requires-Dist: jira<4,>=3.5.0; extra == "jira"
 Provides-Extra: test
 Requires-Dist: pytest<9,>=7.4.0; extra == "test"
+Requires-Dist: pytest-cov>=4.0; extra == "test"
 Requires-Dist: requests<3,>=2.31.0; extra == "test"
 Dynamic: license-file
 
@@ -92,11 +93,12 @@ Requires Python 3.11 or later. Models are downloaded automatically on first run
 
 ```bash
 pip install chunksilo
-
-# Or with Confluence and Jira support:
-pip install chunksilo[confluence,jira]
 ```
 
+Confluence and Jira support is included by default — just provide a config file
+to enable them. (`pip install chunksilo[confluence,jira]` still works as an alias
+for backward compatibility.)
+
 Then:
 1. **Create** a config file at `~/.config/chunksilo/config.yaml` (see [Configuration](#configuration))
 2. **Build** the index: `chunksilo --build-index`
@@ -225,7 +227,7 @@ All settings are optional and have sensible defaults.
 
 #### Confluence Settings (optional)
 
-> **Note:** Confluence integration requires the optional dependency. Install with: `pip install chunksilo[confluence]`
+> **Note:** Confluence support is installed by default; just set the values below to enable it.
 
 | Setting | Description |
 | :--- | :--- |
@@ -243,7 +245,7 @@ All settings are optional and have sensible defaults.
 
 #### Jira Settings (optional)
 
-> **Note:** Jira integration requires the optional dependency. Install with: `pip install chunksilo[jira]`
+> **Note:** Jira support is installed by default; just set the values below to enable it.
 
 | Setting | Description |
 | :--- | :--- |
@@ -418,8 +420,8 @@ Add to `mcp_settings.json` (typically in `~/.config/Code/User/globalStorage/roov
 - **Index missing**: Run `chunksilo --build-index` (PyPI install) or `./venv/bin/chunksilo --build-index` (offline bundle).
 - **Retrieval errors**: Check paths in your MCP client configuration.
 - **Offline mode**: PyPI installs default to `offline: false` (models auto-download). The offline bundle includes pre-downloaded models and sets `offline: true`. Set `retrieval.offline: true` in `config.yaml` to prevent network calls after initial model download.
-- **Confluence Integration**: Install with `pip install chunksilo[confluence]`, then set `confluence.url`, `confluence.username`, and `confluence.api_token` in `config.yaml`.
-- **Jira Integration**: Install with `pip install chunksilo[jira]`, then set `jira.url`, `jira.username`, and `jira.api_token` in `config.yaml`. Optionally configure `jira.projects` to restrict search to specific project keys.
+- **Confluence Integration**: Included by default — set `confluence.url`, `confluence.username`, and `confluence.api_token` in `config.yaml`.
+- **Jira Integration**: Included by default — set `jira.url`, `jira.username`, and `jira.api_token` in `config.yaml`. Optionally configure `jira.projects` to restrict search to specific project keys.
 - **Custom CA Bundle**: Set `ssl.ca_bundle_path` in `config.yaml` for custom certificates.
 - **Network mounts**: Unavailable directories are skipped with a warning; indexing continues with available directories.
 - **Legacy .doc files**: Requires LibreOffice to be installed for automatic conversion to .docx. If LibreOffice is not found, .doc files are skipped with a warning. Full heading extraction is supported.

{chunksilo-2.3.3 → chunksilo-2.4.0}/README.md

@@ -51,11 +51,12 @@ Requires Python 3.11 or later. Models are downloaded automatically on first run
 
 ```bash
 pip install chunksilo
-
-# Or with Confluence and Jira support:
-pip install chunksilo[confluence,jira]
 ```
 
+Confluence and Jira support is included by default — just provide a config file
+to enable them. (`pip install chunksilo[confluence,jira]` still works as an alias
+for backward compatibility.)
+
 Then:
 1. **Create** a config file at `~/.config/chunksilo/config.yaml` (see [Configuration](#configuration))
 2. **Build** the index: `chunksilo --build-index`
@@ -184,7 +185,7 @@ All settings are optional and have sensible defaults.
 
 #### Confluence Settings (optional)
 
-> **Note:** Confluence integration requires the optional dependency. Install with: `pip install chunksilo[confluence]`
+> **Note:** Confluence support is installed by default; just set the values below to enable it.
 
 | Setting | Description |
 | :--- | :--- |
@@ -202,7 +203,7 @@ All settings are optional and have sensible defaults.
 
 #### Jira Settings (optional)
 
-> **Note:** Jira integration requires the optional dependency. Install with: `pip install chunksilo[jira]`
+> **Note:** Jira support is installed by default; just set the values below to enable it.
 
 | Setting | Description |
 | :--- | :--- |
@@ -377,8 +378,8 @@ Add to `mcp_settings.json` (typically in `~/.config/Code/User/globalStorage/roov
 - **Index missing**: Run `chunksilo --build-index` (PyPI install) or `./venv/bin/chunksilo --build-index` (offline bundle).
 - **Retrieval errors**: Check paths in your MCP client configuration.
 - **Offline mode**: PyPI installs default to `offline: false` (models auto-download). The offline bundle includes pre-downloaded models and sets `offline: true`. Set `retrieval.offline: true` in `config.yaml` to prevent network calls after initial model download.
-- **Confluence Integration**: Install with `pip install chunksilo[confluence]`, then set `confluence.url`, `confluence.username`, and `confluence.api_token` in `config.yaml`.
-- **Jira Integration**: Install with `pip install chunksilo[jira]`, then set `jira.url`, `jira.username`, and `jira.api_token` in `config.yaml`. Optionally configure `jira.projects` to restrict search to specific project keys.
+- **Confluence Integration**: Included by default — set `confluence.url`, `confluence.username`, and `confluence.api_token` in `config.yaml`.
+- **Jira Integration**: Included by default — set `jira.url`, `jira.username`, and `jira.api_token` in `config.yaml`. Optionally configure `jira.projects` to restrict search to specific project keys.
 - **Custom CA Bundle**: Set `ssl.ca_bundle_path` in `config.yaml` for custom certificates.
 - **Network mounts**: Unavailable directories are skipped with a warning; indexing continues with available directories.
 - **Legacy .doc files**: Requires LibreOffice to be installed for automatic conversion to .docx. If LibreOffice is not found, .doc files are skipped with a warning. Full heading extraction is supported.

{chunksilo-2.3.3 → chunksilo-2.4.0}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "chunksilo"
-version = "2.3.3"
+version = "2.4.0"
 description = "Local RAG-based semantic document search with MCP server interface"
 license = "Apache-2.0"
 requires-python = ">=3.11"
@@ -25,9 +25,12 @@ classifiers = [
 dynamic = ["dependencies"]
 
 [project.optional-dependencies]
-confluence = ["llama-index-readers-confluence>=0.6.0,<1"]
-jira = ["jira>=3.5.0,<4"]
-test = ["pytest>=7.4.0,<9", "requests>=2.31.0,<3"]
+# Confluence and Jira support is now installed by default (see requirements.txt).
+# These extras are kept as empty aliases so existing `pip install chunksilo[confluence,jira]`
+# commands keep working without a warning.
+confluence = []
+jira = []
+test = ["pytest>=7.4.0,<9", "pytest-cov>=4.0", "requests>=2.31.0,<3"]
 
 [project.scripts]
 chunksilo = "chunksilo.cli:main"
@@ -42,3 +45,24 @@ dependencies = {file = ["requirements.txt"]}
 
 [tool.setuptools.packages.find]
 where = ["src"]
+
+# Bundle the retrieval models into the wheel so air-gapped installs work with no
+# network access. Populated at build time (see the publish job / prepare-common.sh).
+# The HuggingFace `blobs/` copies are excluded: with HF_HUB_DISABLE_SYMLINKS=1 the
+# `snapshots/` tree holds real files, so blobs would only duplicate the bytes.
+[tool.setuptools.package-data]
+chunksilo = ["_bundled_models/**/*"]
+
+[tool.setuptools.exclude-package-data]
+chunksilo = ["_bundled_models/**/blobs/**"]
+
+[tool.ruff]
+target-version = "py311"
+line-length = 120
+
+[tool.ruff.lint]
+select = ["E", "F", "I", "UP"]
+ignore = ["E501"]
+
+[tool.ruff.lint.isort]
+known-first-party = ["chunksilo"]

{chunksilo-2.3.3 → chunksilo-2.4.0}/requirements.txt

@@ -10,3 +10,5 @@ huggingface-hub>=0.22.0,<2
 flashrank>=0.1.0,<1
 fastembed>=0.5.0,<1
 pyyaml>=6.0,<7
+llama-index-readers-confluence>=0.6.0,<1
+jira>=3.5.0,<4

{chunksilo-2.3.3 → chunksilo-2.4.0}/src/chunksilo/__init__.py

@@ -1,4 +1,4 @@
 # SPDX-License-Identifier: Apache-2.0
 """ChunkSilo - Local RAG-based semantic document search."""
 
-__version__ = "2.3.3"
+__version__ = "2.4.0"

{chunksilo-2.3.3 → chunksilo-2.4.0}/src/chunksilo/cfgload.py

@@ -5,14 +5,20 @@ Shared configuration loading for ChunkSilo.
 
 Loads configuration from config.yaml, searching in standard locations.
 """
+import copy
 import logging
 import os
-import yaml
 from pathlib import Path
 from typing import Any
 
+import yaml
+
 logger = logging.getLogger(__name__)
 
+# Retrieval models bundled inside the installed package (populated at wheel-build
+# time). Present in air-gapped wheel installs; absent in plain source checkouts.
+_BUNDLED_MODELS_DIR = Path(__file__).resolve().parent / "_bundled_models"
+
 
 def _find_config() -> Path:
     """Find config.yaml using a priority-based search.
@@ -136,6 +142,46 @@ def _deep_merge(base: dict[str, Any], override: dict[str, Any]) -> dict[str, Any
     return result
 
 
+def _has_embedding_cache(cache_dir: Path) -> bool:
+    """True if ``cache_dir`` holds a HuggingFace-style embedding model cache."""
+    try:
+        return any(cache_dir.glob("models--*"))
+    except OSError:
+        return False
+
+
+def _resolve_bundled_models(config: dict[str, Any]) -> dict[str, Any]:
+    """Transparently fall back to models bundled in the package.
+
+    If the wheel ships the retrieval models and the configured cache directory
+    does not already contain them, point the cache at the bundled copy and enable
+    offline mode. This lets a pip-installed wheel run in an air-gapped environment
+    with zero extra configuration. A user-configured cache that already holds the
+    models always wins, so explicit setups are never overridden.
+
+    Detection looks for any HuggingFace-style ``models--*`` cache directory rather
+    than matching the configured embedding name: fastembed stores the model under
+    its upstream source repo (e.g. ``models--qdrant--bge-small-en-v1.5-onnx-q``),
+    not under the configured alias ``BAAI/bge-small-en-v1.5``.
+    """
+    if not _has_embedding_cache(_BUNDLED_MODELS_DIR):
+        return config  # No models bundled in this install.
+
+    configured = Path(config["storage"]["model_cache_dir"]).expanduser()
+    if configured.resolve() != _BUNDLED_MODELS_DIR.resolve() and _has_embedding_cache(
+        configured
+    ):
+        return config  # Configured cache already has models; respect it.
+
+    result = copy.deepcopy(config)
+    result["storage"]["model_cache_dir"] = str(_BUNDLED_MODELS_DIR)
+    result["retrieval"]["offline"] = True
+    logger.info(
+        "Using bundled retrieval models at %s (offline mode)", _BUNDLED_MODELS_DIR
+    )
+    return result
+
+
 def load_config(config_path: Path | None = None) -> dict[str, Any]:
     """Load configuration from YAML file with defaults.
 
@@ -156,14 +202,15 @@ def load_config(config_path: Path | None = None) -> dict[str, Any]:
 
     if not path.exists():
         logger.info("Config file not found at %s; using built-in defaults", path)
-        return _DEFAULTS.copy()
+        return _resolve_bundled_models(_DEFAULTS.copy())
 
     logger.info("Using config: %s", path)
 
-    with open(path, "r", encoding="utf-8") as f:
+    with open(path, encoding="utf-8") as f:
         user_config = yaml.safe_load(f) or {}
 
     result = _deep_merge(_DEFAULTS, user_config)
+    result = _resolve_bundled_models(result)
 
     # Cache result only for default path
     if config_path is None:

{chunksilo-2.3.3 → chunksilo-2.4.0}/src/chunksilo/cli.py

@@ -57,8 +57,9 @@ def main():
     logging.basicConfig(level=log_level, format="%(message)s", stream=sys.stderr)
 
     if args.dump_defaults:
-        from .cfgload import _DEFAULTS
         import yaml
+
+        from .cfgload import _DEFAULTS
         yaml.dump(_DEFAULTS, sys.stdout, default_flow_style=False, sort_keys=False)
         return
 

chunksilo-2.4.0/src/chunksilo/docx_utils.py

@@ -0,0 +1,285 @@
+#!/usr/bin/env python3
+# SPDX-License-Identifier: Apache-2.0
+"""DOCX/DOC document processing utilities.
+
+Handles parsing DOCX files into heading-based documents, heading level
+extraction, and .doc-to-.docx conversion via LibreOffice.
+"""
+from __future__ import annotations
+
+import logging
+from datetime import datetime
+from pathlib import Path
+from typing import TYPE_CHECKING, Any
+
+from docx import Document
+from llama_index.core import Document as LlamaIndexDocument
+
+from . import cfgload
+
+if TYPE_CHECKING:
+    from .ui import FileProcessingContext
+
+logger = logging.getLogger(__name__)
+
+
+def _parse_heading_level(style_name: str | None) -> int:
+    """Best-effort extraction of a numeric heading level from a DOCX style name."""
+    if not style_name:
+        return 1
+    try:
+        if "Heading" in style_name:
+            level_str = style_name.replace("Heading", "").strip()
+            if level_str:
+                return int(level_str)
+    except (ValueError, AttributeError):
+        pass
+    return 1
+
+
+def _is_heading_style(style_name: str) -> bool:
+    """Check whether a DOCX style name represents a heading."""
+    return (
+        style_name.startswith("Heading")
+        or style_name.startswith("heading")
+        or "Heading" in style_name
+    )
+
+
+def _get_doc_temp_dir() -> Path:
+    """Get the temporary directory for .doc conversion, creating it if needed."""
+    storage_dir = Path(cfgload.get("storage.storage_dir", "./storage"))
+    temp_dir = storage_dir / "doc_temp"
+    temp_dir.mkdir(parents=True, exist_ok=True)
+    return temp_dir
+
+
+def _convert_doc_to_docx(doc_path: Path, timeout: float = 60) -> Path | None:
+    """Convert a .doc file to .docx using LibreOffice.
+
+    Args:
+        doc_path: Path to .doc file
+        timeout: Timeout in seconds for conversion process
+
+    Returns:
+        Path to temporary .docx file, or None if conversion fails.
+        Caller is responsible for cleaning up the temp file.
+    """
+    import shutil
+    import subprocess
+
+    # Find LibreOffice executable
+    soffice_paths = [
+        "/Applications/LibreOffice.app/Contents/MacOS/soffice",  # macOS
+        "/usr/bin/soffice",  # Linux
+        "/usr/bin/libreoffice",  # Linux alternative
+        "soffice",  # Windows (in PATH)
+    ]
+
+    soffice = None
+    for path in soffice_paths:
+        if shutil.which(path):
+            soffice = path
+            break
+
+    if not soffice:
+        logger.warning(f"LibreOffice not found. Cannot convert {doc_path}")
+        return None
+
+    # Use storage directory for temp files (more reliable space than /tmp)
+    temp_dir = _get_doc_temp_dir()
+
+    try:
+        result = subprocess.run(
+            [soffice, "--headless", "--convert-to", "docx",
+             "--outdir", str(temp_dir), str(doc_path)],
+            capture_output=True,
+            timeout=timeout,
+        )
+        if result.returncode != 0:
+            logger.warning(f"LibreOffice conversion failed for {doc_path}: {result.stderr}")
+            return None
+
+        # Find the converted file
+        docx_name = doc_path.stem + ".docx"
+        docx_path = temp_dir / docx_name
+        if docx_path.exists():
+            return docx_path
+
+        logger.warning(f"Converted file not found: {docx_path}")
+    except subprocess.TimeoutExpired:
+        logger.warning(f"LibreOffice conversion timed out for {doc_path}")
+    except Exception as e:
+        logger.warning(f"Error converting {doc_path}: {e}")
+
+    return None
+
+
+def split_docx_into_heading_documents(
+    docx_path: Path,
+    ctx: FileProcessingContext | None = None,
+    *,
+    heading_store: Any = None,
+    excluded_embed_metadata_keys: list[str] | None = None,
+    excluded_llm_metadata_keys: list[str] | None = None,
+) -> list[LlamaIndexDocument]:
+    """Split DOCX into documents by heading with progress updates.
+
+    Args:
+        docx_path: Path to DOCX file
+        ctx: Optional processing context for progress updates and timeout
+        heading_store: HeadingStore instance for persisting heading metadata
+        excluded_embed_metadata_keys: Keys to exclude from embedding text
+        excluded_llm_metadata_keys: Keys to exclude from LLM context
+    """
+    _excluded_embed = excluded_embed_metadata_keys or []
+    _excluded_llm = excluded_llm_metadata_keys or []
+
+    docs: list[LlamaIndexDocument] = []
+
+    if ctx:
+        ctx.set_phase("Opening DOCX")
+
+    try:
+        doc = Document(docx_path)
+    except Exception as e:
+        logger.warning(f"Failed to open DOCX {docx_path}: {e}")
+        return docs
+
+    # Extract file dates from filesystem
+    stat = docx_path.stat()
+    creation_date = datetime.fromtimestamp(stat.st_ctime).strftime("%Y-%m-%d")
+    last_modified_date = datetime.fromtimestamp(stat.st_mtime).strftime("%Y-%m-%d")
+
+    # Try to extract dates from DOCX core properties (more accurate than filesystem)
+    try:
+        core_props = doc.core_properties
+        if core_props.created:
+            creation_date = core_props.created.strftime("%Y-%m-%d")
+        if core_props.modified:
+            last_modified_date = core_props.modified.strftime("%Y-%m-%d")
+    except Exception:
+        logger.debug("Could not read DOCX core properties for %s, using filesystem dates", docx_path, exc_info=True)
+
+    # First pass: Extract all headings with positions for hierarchy metadata
+    if ctx:
+        ctx.set_phase("Extracting headings")
+
+    all_headings = []
+    char_position = 0
+    for para in doc.paragraphs:
+        # Periodically check for timeout during long operations
+        if ctx and len(all_headings) % 100 == 0:
+            ctx.set_phase(f"Extracting headings ({len(all_headings)} found)")
+
+        style_name = getattr(para.style, "name", "") or ""
+
+        if _is_heading_style(style_name) and para.text.strip():
+            heading_level = _parse_heading_level(style_name)
+            all_headings.append({
+                "text": para.text.strip(),
+                "position": char_position,
+                "level": heading_level
+            })
+
+        char_position += len(para.text) + 1  # +1 for newline
+
+    # Store headings separately to avoid metadata size issues during chunking
+    if heading_store is not None:
+        if ctx:
+            ctx.set_phase("Storing heading metadata")
+        heading_store.set_headings(str(docx_path), all_headings)
+
+    # Second pass: Split by heading (existing logic)
+    if ctx:
+        ctx.set_phase("Splitting into sections")
+    current_heading: str | None = None
+    current_level: int | None = None
+    current_body: list[str] = []
+
+    def flush_current():
+        if not current_heading:
+            return
+        text = "\n".join(line for line in current_body if line is not None).strip()
+        if not text:
+            return
+
+        # Build hierarchical heading_path by finding parent headings based on level
+        heading_path = []
+        if all_headings:
+            # Find the index of the current heading in all_headings
+            current_idx = None
+            for idx, h in enumerate(all_headings):
+                if h["text"] == current_heading and h["level"] == current_level:
+                    current_idx = idx
+                    break
+
+            if current_idx is not None:
+                # Build path by including all parent headings (those with lower level numbers)
+                # Walk backwards from current heading and include headings with level < current_level
+                path_headings = [all_headings[current_idx]]  # Start with current
+                for idx in range(current_idx - 1, -1, -1):
+                    h = all_headings[idx]
+                    if h["level"] < path_headings[0]["level"]:
+                        path_headings.insert(0, h)
+                heading_path = [h["text"] for h in path_headings]
+
+        metadata = {
+            "file_path": str(docx_path),
+            "file_name": docx_path.name,
+            "source": str(docx_path),
+            "heading": current_heading,
+            "heading_level": current_level,
+            "creation_date": creation_date,
+            "last_modified_date": last_modified_date,
+            "heading_path": heading_path,  # Pre-computed hierarchical path
+        }
+        docs.append(LlamaIndexDocument(
+            text=text,
+            metadata=metadata,
+            excluded_embed_metadata_keys=_excluded_embed,
+            excluded_llm_metadata_keys=_excluded_llm,
+        ))
+
+    for para in doc.paragraphs:
+        style_name = getattr(para.style, "name", "") or ""
+
+        if _is_heading_style(style_name) and para.text.strip():
+            flush_current()
+            current_heading = para.text.strip()
+            current_level = _parse_heading_level(style_name)
+            current_body = []
+        else:
+            if current_heading is not None:
+                current_body.append(para.text)
+
+    flush_current()
+
+    if not docs:
+        try:
+            full_text = "\n".join(p.text for p in doc.paragraphs).strip()
+        except Exception:
+            logger.debug("Failed to extract paragraph text from %s", docx_path, exc_info=True)
+            full_text = ""
+
+        if full_text:
+            metadata = {
+                "file_path": str(docx_path),
+                "file_name": docx_path.name,
+                "source": str(docx_path),
+                "heading": None,
+                "heading_level": None,
+                "creation_date": creation_date,
+                "last_modified_date": last_modified_date,
+            }
+            docs.append(LlamaIndexDocument(
+                text=full_text,
+                metadata=metadata,
+                excluded_embed_metadata_keys=_excluded_embed,
+                excluded_llm_metadata_keys=_excluded_llm,
+            ))
+
+    logger.info(
+        f"Split DOCX {docx_path} into {len(docs)} heading-based document(s)"
+    )
+    return docs