scitex 2.15.3__py3-none-any.whl → 2.16.0__py3-none-any.whl

scitex/scholar/_mcp/openalex_tool_schemas.py

@@ -0,0 +1,96 @@
+#!/usr/bin/env python3
+# Timestamp: 2026-01-29
+# File: src/scitex/scholar/_mcp/openalex_tool_schemas.py
+"""OpenAlex-Local tool schemas for MCP server.
+
+Provides access to 250M+ papers via openalex-local database.
+"""
+
+from __future__ import annotations
+
+import mcp.types as types
+
+__all__ = ["get_openalex_tool_schemas"]
+
+
+def get_openalex_tool_schemas() -> list[types.Tool]:
+    """Return OpenAlex-Local tool schemas."""
+    return [
+        types.Tool(
+            name="openalex_search",
+            description=(
+                "Search OpenAlex database (250M+ papers) via openalex-local. "
+                "Fast full-text search with year filtering. Includes citation data."
+            ),
+            inputSchema={
+                "type": "object",
+                "properties": {
+                    "query": {
+                        "type": "string",
+                        "description": "Search query string (full-text search)",
+                    },
+                    "limit": {
+                        "type": "integer",
+                        "description": "Maximum number of results",
+                        "default": 20,
+                    },
+                    "offset": {
+                        "type": "integer",
+                        "description": "Number of results to skip for pagination",
+                        "default": 0,
+                    },
+                    "year_min": {
+                        "type": "integer",
+                        "description": "Minimum publication year filter",
+                    },
+                    "year_max": {
+                        "type": "integer",
+                        "description": "Maximum publication year filter",
+                    },
+                },
+                "required": ["query"],
+            },
+        ),
+        types.Tool(
+            name="openalex_get",
+            description="Get a paper by DOI or OpenAlex ID from OpenAlex database.",
+            inputSchema={
+                "type": "object",
+                "properties": {
+                    "doi": {
+                        "type": "string",
+                        "description": "DOI of the paper (e.g., '10.1038/nature12373')",
+                    },
+                    "openalex_id": {
+                        "type": "string",
+                        "description": "OpenAlex ID (e.g., 'W2100837269')",
+                    },
+                },
+            },
+        ),
+        types.Tool(
+            name="openalex_count",
+            description="Count papers matching a search query in OpenAlex database.",
+            inputSchema={
+                "type": "object",
+                "properties": {
+                    "query": {
+                        "type": "string",
+                        "description": "Search query string",
+                    },
+                },
+                "required": ["query"],
+            },
+        ),
+        types.Tool(
+            name="openalex_info",
+            description="Get OpenAlex database configuration and status.",
+            inputSchema={
+                "type": "object",
+                "properties": {},
+            },
+        ),
+    ]
+
+
+# EOF

scitex/scholar/_mcp/tool_schemas.py

@@ -3,9 +3,14 @@
 # File: src/scitex/scholar/_mcp.tool_schemas.py
 # ----------------------------------------
 """Tool schemas for the scitex-scholar MCP server."""
+
 from __future__ import annotations
+
 import mcp.types as types
+
 __all__ = ["get_tool_schemas"]
+
+
 def get_tool_schemas() -> list[types.Tool]:
     """Return all tool schemas for the Scholar MCP server."""
     base_schemas = [
@@ -500,6 +505,16 @@ def get_tool_schemas() -> list[types.Tool]:
             },
         ),
     ]
+    from .crossref_tool_schemas import get_crossref_tool_schemas
     from .job_tool_schemas import get_job_tool_schemas
-    return base_schemas + get_job_tool_schemas()
+    from .openalex_tool_schemas import get_openalex_tool_schemas
+
+    return (
+        base_schemas
+        + get_job_tool_schemas()
+        + get_crossref_tool_schemas()
+        + get_openalex_tool_schemas()
+    )
+
+
 # EOF

scitex/scholar/examples/03_02-engine-for-bibtex.py

@@ -54,7 +54,7 @@ async def search_bibtex_metadata(
 ) -> list:
     """Search for metadata of papers from BibTeX file."""
     from scitex.scholar import ScholarConfig, ScholarEngine
-    from scitex.scholar.utils import parse_bibtex
+    from scitex.scholar._utils import parse_bibtex
 
     entries = parse_bibtex(bibtex_path)
     if n_samples:

scitex/scholar/examples/04_02-url-for-bibtex.py

@@ -83,7 +83,7 @@ async def process_bibtex_urls(
         ScholarEngine,
         ScholarURLFinder,
     )
-    from scitex.scholar.utils import parse_bibtex
+    from scitex.scholar._utils import parse_bibtex
 
     print(f"📚 Parsing BibTeX file: {bibtex_path}")
     entries = parse_bibtex(bibtex_path)

scitex/scholar/examples/06_parse_bibtex.py

@@ -66,7 +66,7 @@ def demonstrate_bibtex_parsing(
     dict
         Parsed BibTeX entries
     """
-    from scitex.scholar.utils import parse_bibtex
+    from scitex.scholar._utils import parse_bibtex
 
     results = {}
 

scitex/scholar/examples/99_fullpipeline-for-bibtex.py

@@ -85,7 +85,7 @@ async def process_bibtex_entries(
         ScholarPDFDownloader,
         ScholarURLFinder,
     )
-    from scitex.scholar.utils import parse_bibtex
+    from scitex.scholar._utils import parse_bibtex
 
     # Parse BibTeX entries
     entries = parse_bibtex(bibtex_path)

scitex/scholar/local_dbs/__init__.py

@@ -0,0 +1,31 @@
+#!/usr/bin/env python3
+# Timestamp: 2026-01-29
+# File: src/scitex/scholar/local_dbs/__init__.py
+"""Local database integrations for scitex.scholar.
+
+This package provides access to local scholarly databases:
+- crossref_scitex: CrossRef database (167M+ papers via crossref-local)
+- openalex_scitex: OpenAlex database (284M+ works via openalex-local)
+
+Both modules delegate directly to their respective external packages
+without any additional logic.
+
+Usage:
+    >>> from scitex.scholar.local_dbs import crossref_scitex
+    >>> results = crossref_scitex.search("machine learning")
+
+    >>> from scitex.scholar.local_dbs import openalex_scitex
+    >>> results = openalex_scitex.search("neural networks")
+"""
+
+from __future__ import annotations
+
+from . import crossref_scitex, openalex_scitex
+
+__all__ = [
+    "crossref_scitex",
+    "openalex_scitex",
+]
+
+
+# EOF

scitex/scholar/local_dbs/crossref_scitex.py

@@ -0,0 +1,30 @@
+#!/usr/bin/env python3
+# Timestamp: 2026-01-29
+# File: src/scitex/scholar/local_dbs/crossref_scitex.py
+"""CrossRef-SciTeX: Minimal API for crossref-local.
+
+Usage:
+    >>> from scitex.scholar.local_dbs import crossref_scitex
+    >>> results = crossref_scitex.search("machine learning")
+    >>> work = crossref_scitex.get("10.1038/nature12373")
+"""
+
+try:
+    from crossref_local import (
+        SearchResult,
+        # Classes
+        Work,
+        count,
+        get,
+        info,
+        # Core functions
+        search,
+    )
+except ImportError as e:
+    raise ImportError(
+        "crossref-local not installed. Install with: pip install crossref-local"
+    ) from e
+
+__all__ = ["search", "get", "count", "info", "Work", "SearchResult"]
+
+# EOF

scitex/scholar/local_dbs/openalex_scitex.py

@@ -0,0 +1,30 @@
+#!/usr/bin/env python3
+# Timestamp: 2026-01-29
+# File: src/scitex/scholar/local_dbs/openalex_scitex.py
+"""OpenAlex-SciTeX: Minimal API for openalex-local.
+
+Usage:
+    >>> from scitex.scholar.local_dbs import openalex_scitex
+    >>> results = openalex_scitex.search("machine learning")
+    >>> work = openalex_scitex.get("10.1038/nature12373")
+"""
+
+try:
+    from openalex_local import (
+        SearchResult,
+        # Classes
+        Work,
+        count,
+        get,
+        info,
+        # Core functions
+        search,
+    )
+except ImportError as e:
+    raise ImportError(
+        "openalex-local not installed. Install with: pip install openalex-local"
+    ) from e
+
+__all__ = ["search", "get", "count", "info", "Work", "SearchResult"]
+
+# EOF

scitex/scholar/mcp_server.py

@@ -3,8 +3,17 @@
 # File: src/scitex/scholar/mcp_server.py
 # ----------------------------------------
 
-"""
-MCP Server for SciTeX Scholar - Scientific Literature Management
+"""MCP Server for SciTeX Scholar - Scientific Literature Management.
+
+.. deprecated::
+    This standalone server is deprecated. Use the unified scitex MCP server instead:
+
+    CLI: scitex serve
+    Python: from scitex.mcp_server import run_server
+
+    The unified server includes all scholar tools plus other scitex tools.
+    Scholar tools are prefixed with 'scholar_' (e.g., scholar_search_papers).
+    Scholar resources are available at scholar://library and scholar://bibtex.
 
 Provides tools for:
 - Searching papers across multiple databases
@@ -16,6 +25,15 @@ Provides tools for:
 
 from __future__ import annotations
 
+import warnings
+
+warnings.warn(
+    "scitex.scholar.mcp_server is deprecated. Use 'scitex serve' or "
+    "'from scitex.mcp_server import run_server' for the unified MCP server.",
+    DeprecationWarning,
+    stacklevel=2,
+)
+
 import asyncio
 import os
 from datetime import datetime
@@ -67,11 +85,18 @@ class ScholarServer:
 
                 self._scholar_instance = Scholar()
             except ImportError as e:
-                raise RuntimeError(f"Scholar module not available: {e}")
+                raise RuntimeError(f"Scholar module not available: {e}") from e
         return self._scholar_instance
 
     def setup_handlers(self):
         """Set up MCP server handlers."""
+        from ._mcp.crossref_handlers import (
+            crossref_citations_handler,
+            crossref_count_handler,
+            crossref_get_handler,
+            crossref_info_handler,
+            crossref_search_handler,
+        )
         from ._mcp.handlers import (
             add_papers_to_project_handler,
             authenticate_handler,
@@ -186,6 +211,36 @@ class ScholarServer:
             elif name == "get_job_result":
                 return await self._wrap_result(get_job_result_handler(**arguments))
 
+            # CrossRef-Local Tools
+            elif name == "crossref_search":
+                return await self._wrap_result(crossref_search_handler(**arguments))
+            elif name == "crossref_get":
+                return await self._wrap_result(crossref_get_handler(**arguments))
+            elif name == "crossref_count":
+                return await self._wrap_result(crossref_count_handler(**arguments))
+            elif name == "crossref_citations":
+                return await self._wrap_result(crossref_citations_handler(**arguments))
+            elif name == "crossref_info":
+                return await self._wrap_result(crossref_info_handler(**arguments))
+
+            # OpenAlex-Local Tools
+            elif name == "openalex_search":
+                from ._mcp.openalex_handlers import openalex_search_handler
+
+                return await self._wrap_result(openalex_search_handler(**arguments))
+            elif name == "openalex_get":
+                from ._mcp.openalex_handlers import openalex_get_handler
+
+                return await self._wrap_result(openalex_get_handler(**arguments))
+            elif name == "openalex_count":
+                from ._mcp.openalex_handlers import openalex_count_handler
+
+                return await self._wrap_result(openalex_count_handler(**arguments))
+            elif name == "openalex_info":
+                from ._mcp.openalex_handlers import openalex_info_handler
+
+                return await self._wrap_result(openalex_info_handler(**arguments))
+
             else:
                 raise ValueError(f"Unknown tool: {name}")
 
@@ -340,7 +395,7 @@ async def _run_server():
 
 
 def main():
-    """Main entry point for the MCP server."""
+    """Run the MCP server."""
     if not MCP_AVAILABLE:
         import sys
 

scitex/scholar/pipelines/ScholarPipelineMetadataSingle.py

@@ -32,7 +32,7 @@ from scitex.scholar.core import Paper
 from scitex.scholar.metadata_engines import ScholarEngine
 from scitex.scholar.impact_factor import ImpactFactorEngine
 from scitex.scholar.config import ScholarConfig
-from scitex.scholar.utils import DOIValidator
+from scitex.scholar._utils import DOIValidator
 
 logger = logging.getLogger(__name__)
 

scitex/scholar/storage/_mixins/_paper_saving.py

@@ -15,7 +15,7 @@ from datetime import datetime
 from typing import TYPE_CHECKING, Any, Dict, List, Optional
 
 from scitex import logging
-from scitex.scholar.utils import TextNormalizer
+from scitex.scholar._utils import TextNormalizer
 
 if TYPE_CHECKING:
     from scitex.scholar.core.Paper import Paper

scitex/scholar/storage/_mixins/_resolution.py

@@ -14,7 +14,7 @@ from datetime import datetime
 from typing import Any, Dict, List, Optional
 
 from scitex import logging
-from scitex.scholar.utils import TextNormalizer
+from scitex.scholar._utils import TextNormalizer
 
 logger = logging.getLogger(__name__)
 

scitex/scholar/url_finder/strategies/find_pdf_urls_by_zotero_translators.py

@@ -1,9 +1,9 @@
 #!/usr/bin/env python3
-# -*- coding: utf-8 -*-
 # Timestamp: "2025-10-13 06:32:08 (ywatanabe)"
 # File: /home/ywatanabe/proj/scitex_repo/src/scitex/scholar/url_finder/strategies/find_pdf_urls_by_zotero_translators.py
 # ----------------------------------------
 from __future__ import annotations
+
 import os
 
 __FILE__ = (
@@ -38,10 +38,10 @@ Usage:
 from typing import List
 
 from playwright.async_api import Page
-from scitex.scholar.url_finder.translators.core.registry import TranslatorRegistry
 
 from scitex import logging
 from scitex.browser import browser_logger
+from scitex.scholar.url_finder.translators._core.registry import TranslatorRegistry
 
 logger = logging.getLogger(__name__)
 
@@ -67,11 +67,13 @@ async def find_pdf_urls_by_zotero_translators(
         config: ScholarConfig instance (unused, for signature consistency)
         func_name: Function name for logging
 
-    Returns:
+    Returns
+    -------
         List of PDF URLs extracted by matching translators
         Empty list if no translator matches or extraction fails
 
-    Examples:
+    Examples
+    --------
         >>> async with async_playwright() as p:
         ...     browser = await p.chromium.launch()
         ...     page = await browser.new_page()

scitex/scholar/url_finder/translators/__init__.py

@@ -21,11 +21,12 @@ Internal modules (not part of public API):
 """
 
 # Public API - only expose what users need
-from .core.registry import TranslatorRegistry
-from .core.base import BaseTranslator
+from ._core.base import (
+    BaseTranslator as _BaseTranslator,
+)  # Hidden, use TranslatorRegistry
+from ._core.registry import TranslatorRegistry
 
 # Everything else is internal implementation
 __all__ = [
     "TranslatorRegistry",
-    "BaseTranslator",
 ]