langparse 0.1.0__tar.gz → 0.1.0rc2__tar.gz

{langparse-0.1.0/langparse.egg-info → langparse-0.1.0rc2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: langparse
-Version: 0.1.0
+Version: 0.1.0rc2
 Summary: A developer-friendly document parsing toolkit with precise, source-grounded Excel understanding.
 Author-email: syw2014 <syw2014@gmail.com>
 License-Expression: Apache-2.0
@@ -40,7 +40,7 @@ Provides-Extra: deepdoc
 Requires-Dist: pdfplumber>=0.11.10; extra == "deepdoc"
 Requires-Dist: opencv-python-headless>=4.9.0; extra == "deepdoc"
 Requires-Dist: onnxruntime>=1.17.0; extra == "deepdoc"
-Requires-Dist: pypdf>=6.19.0; extra == "deepdoc"
+Requires-Dist: pypdf>=6.15.0; extra == "deepdoc"
 Requires-Dist: huggingface_hub>=0.20.0; extra == "deepdoc"
 Requires-Dist: scikit-learn>=1.3.0; extra == "deepdoc"
 Requires-Dist: shapely>=2.0.0; extra == "deepdoc"
@@ -58,7 +58,7 @@ Requires-Dist: rapidocr_onnxruntime; extra == "all"
 Requires-Dist: mineru<4,>=3.4; extra == "all"
 Requires-Dist: opencv-python-headless; extra == "all"
 Requires-Dist: onnxruntime; extra == "all"
-Requires-Dist: pypdf>=6.19.0; extra == "all"
+Requires-Dist: pypdf>=6.15.0; extra == "all"
 Requires-Dist: huggingface_hub; extra == "all"
 Requires-Dist: scikit-learn; extra == "all"
 Requires-Dist: shapely; extra == "all"
@@ -112,59 +112,13 @@ where LangParse goes deliberately deeper.
 
 ---
 
-## Parsing progress (unreleased)
-
-Pass `progress_callback` to `AutoParser`, `ParseService`, `PDFParser`, or
-`BatchParseService.run`. The callback receives immutable `ProgressEvent` objects:
-
-```python
-from langparse import AutoParser
-
-result = AutoParser.parse_result(
-    "budget.xlsx",
-    progress_callback=lambda event: print(event.phase, event.state, event.percent),
-)
-```
-
-Events contain `source`, `phase`, `state`, `completed_units`, `total_units`, `unit`,
-`percent`, and `message`. Percent is phase-local (0–100), not an ETA; unknown
-values remain `None`. Only a `file` event with `completed`/`failed` marks the end
-of parsing and requested chunking, before any caller-side export. A completed
-parse can still have partial quality diagnostics.
-
-Simple PDF reports page counts; DeepDoc forwards its internal weighted-stage or
-page-batch progress; MinerU reports preparation, remote parsing, and rendering
-without an internal percentage. Excel reports extraction, assembly (including
-optional model disambiguation), and rendering. Direct `ExcelParser` calls also
-emit those stages; the facade/service supplies the file lifecycle.
-
-Batch events use `phase="batch"`, `source=""`, and `unit="files"`. Notifications
-follow completion order while returned results retain the batch's sorted file order. Success,
-failure, and skip all count as finished files. Batch completion follows report
-writing and does not mean every file succeeded. Each `batch_item` event names
-the source and reports `completed`/`failed`/`skipped` after that item's output
-write. Empty batches report 0/0 with
-unknown percent. Ordinary callback exceptions are logged by type and isolated
-from parsing. Callbacks are synchronous: keep them brief. One batch serializes
-its callbacks, which can run on different threads; callers coordinate callbacks
-shared across independent runs. There is no task store, background execution,
-cancellation, heartbeat, or HTTP query endpoint.
-
-CLI `--progress` writes line-oriented updates to stderr, leaving stdout intact:
-
-```bash
-langparse parse budget.xlsx --format json --progress > result.json
-langparse parse docs/ --batch --output-dir out --progress
-```
-
 ## Project status
 
-The latest GitHub release is `0.1.0`. Core multi-format parsing,
+The current release candidate is `0.1.0rc2`. Core multi-format parsing,
 structured OOXML workbooks, semantic chunking, batch processing, quality checks,
 and CI are available today. LangParse remains pre-1.0; see
 [docs/PROGRESS.md](docs/PROGRESS.md) for the module-by-module source of truth and
-known gaps. See the
-[release scope](docs/RELEASE_0.1.0.md).
+known gaps.
 
 ## Why LangParse?
 
@@ -214,7 +168,7 @@ views derived from it, not replacements for it.
 Install the current release candidate:
 
 ```bash
-pip install "langparse==0.1.0"
+pip install --pre "langparse==0.1.0rc2"
 ```
 
 Install only the optional capabilities you need:
@@ -782,9 +736,3 @@ See [CHANGELOG.md](CHANGELOG.md) ([中文](CHANGELOG_cn.md)) for release notes a
 
 ## License
 This project is licensed under the [Apache 2.0 License](https://www.apache.org/licenses/LICENSE-2.0).
-
-### v0.1.0 capabilities
-
-- [Workbook Bundle 与查询](docs/WORKBOOK_BUNDLE.md)
-- [分块策略与 CLI](docs/CHUNKING.md)
-- [Agent Skill 与接入示例](SKILLS.md)

{langparse-0.1.0 → langparse-0.1.0rc2}/README.md

@@ -33,59 +33,13 @@ where LangParse goes deliberately deeper.
 
 ---
 
-## Parsing progress (unreleased)
-
-Pass `progress_callback` to `AutoParser`, `ParseService`, `PDFParser`, or
-`BatchParseService.run`. The callback receives immutable `ProgressEvent` objects:
-
-```python
-from langparse import AutoParser
-
-result = AutoParser.parse_result(
-    "budget.xlsx",
-    progress_callback=lambda event: print(event.phase, event.state, event.percent),
-)
-```
-
-Events contain `source`, `phase`, `state`, `completed_units`, `total_units`, `unit`,
-`percent`, and `message`. Percent is phase-local (0–100), not an ETA; unknown
-values remain `None`. Only a `file` event with `completed`/`failed` marks the end
-of parsing and requested chunking, before any caller-side export. A completed
-parse can still have partial quality diagnostics.
-
-Simple PDF reports page counts; DeepDoc forwards its internal weighted-stage or
-page-batch progress; MinerU reports preparation, remote parsing, and rendering
-without an internal percentage. Excel reports extraction, assembly (including
-optional model disambiguation), and rendering. Direct `ExcelParser` calls also
-emit those stages; the facade/service supplies the file lifecycle.
-
-Batch events use `phase="batch"`, `source=""`, and `unit="files"`. Notifications
-follow completion order while returned results retain the batch's sorted file order. Success,
-failure, and skip all count as finished files. Batch completion follows report
-writing and does not mean every file succeeded. Each `batch_item` event names
-the source and reports `completed`/`failed`/`skipped` after that item's output
-write. Empty batches report 0/0 with
-unknown percent. Ordinary callback exceptions are logged by type and isolated
-from parsing. Callbacks are synchronous: keep them brief. One batch serializes
-its callbacks, which can run on different threads; callers coordinate callbacks
-shared across independent runs. There is no task store, background execution,
-cancellation, heartbeat, or HTTP query endpoint.
-
-CLI `--progress` writes line-oriented updates to stderr, leaving stdout intact:
-
-```bash
-langparse parse budget.xlsx --format json --progress > result.json
-langparse parse docs/ --batch --output-dir out --progress
-```
-
 ## Project status
 
-The latest GitHub release is `0.1.0`. Core multi-format parsing,
+The current release candidate is `0.1.0rc2`. Core multi-format parsing,
 structured OOXML workbooks, semantic chunking, batch processing, quality checks,
 and CI are available today. LangParse remains pre-1.0; see
 [docs/PROGRESS.md](docs/PROGRESS.md) for the module-by-module source of truth and
-known gaps. See the
-[release scope](docs/RELEASE_0.1.0.md).
+known gaps.
 
 ## Why LangParse?
 
@@ -135,7 +89,7 @@ views derived from it, not replacements for it.
 Install the current release candidate:
 
 ```bash
-pip install "langparse==0.1.0"
+pip install --pre "langparse==0.1.0rc2"
 ```
 
 Install only the optional capabilities you need:
@@ -703,9 +657,3 @@ See [CHANGELOG.md](CHANGELOG.md) ([中文](CHANGELOG_cn.md)) for release notes a
 
 ## License
 This project is licensed under the [Apache 2.0 License](https://www.apache.org/licenses/LICENSE-2.0).
-
-### v0.1.0 capabilities
-
-- [Workbook Bundle 与查询](docs/WORKBOOK_BUNDLE.md)
-- [分块策略与 CLI](docs/CHUNKING.md)
-- [Agent Skill 与接入示例](SKILLS.md)

{langparse-0.1.0 → langparse-0.1.0rc2}/langparse/__init__.py

@@ -3,14 +3,7 @@ from importlib.metadata import version as _distribution_version
 __version__ = _distribution_version("langparse")
 
 from langparse.autoparser import AutoParser
-from langparse.chunkers import (
-    FixedTokenChunker,
-    SemanticChunker,
-    SlidingWindowChunker,
-    available_chunkers,
-    create_chunker,
-    register_chunker,
-)
+from langparse.chunkers.semantic import SemanticChunker
 from langparse.core.chunker import BaseChunker
 from langparse.core.parser import BaseParser
 from langparse.metrics import BatchItemResult, BatchRunResult, ParseMetrics
@@ -18,7 +11,6 @@ from langparse.parsers.docx_parser import DocxParser
 from langparse.parsers.excel_parser import ExcelParser
 from langparse.parsers.markdown_parser import MarkdownParser
 from langparse.parsers.pdf_parser import PDFParser
-from langparse.progress import ProgressCallback, ProgressEvent
 from langparse.types import (
     Chunk,
     Document,
@@ -42,14 +34,7 @@ __all__ = [
     "DocxParser",
     "ExcelParser",
     "SemanticChunker",
-    "FixedTokenChunker",
-    "SlidingWindowChunker",
-    "available_chunkers",
-    "create_chunker",
-    "register_chunker",
     "ParseMetrics",
-    "ProgressCallback",
-    "ProgressEvent",
     "BatchItemResult",
     "BatchRunResult",
 ]

{langparse-0.1.0 → langparse-0.1.0rc2}/langparse/chunkers/workbook.py

@@ -1,7 +1,6 @@
 from __future__ import annotations
 
 from collections.abc import Callable
-from dataclasses import asdict
 
 from openpyxl.utils import get_column_letter, range_boundaries
 
@@ -19,7 +18,6 @@ from langparse.workbooks.types import (
     LogicalTable,
     MatrixBlock,
     MatrixHeader,
-    SourceRef,
     TableContinuation,
     TextBlock,
     TextLine,
@@ -115,18 +113,6 @@ class WorkbookStructuralChunker:
                 sheet_snapshot.visibility if sheet_snapshot is not None else sheet_ir.visibility
             )
             chunk.metadata["hidden_row_numbers"] = sorted(referenced_rows & hidden_rows)
-            if self.policy.analysis_records:
-                lineage = workbook_ir.lineage
-                dependencies = []
-                for source in chunk.metadata["source_ranges"]:
-                    name, cell_range = source.rsplit("!", 1)
-                    for edge in (
-                        lineage.dependencies_of(SourceRef(name, cell_range)) if lineage else []
-                    ):
-                        payload = asdict(edge)
-                        if payload not in dependencies:
-                            dependencies.append(payload)
-                chunk.structured_payload["dependencies"] = dependencies
 
     def _validate_chunks(self, parsed: ParsedDocumentResult, chunks: list[Chunk]) -> None:
         workbook_ir = parsed.structure
@@ -155,12 +141,7 @@ class WorkbookStructuralChunker:
         for chunk in chunks:
             source_ranges = chunk.metadata["source_ranges"]
             for source_range in source_ranges:
-                if chunk.metadata["chunk_type"] in {"chart", "image"}:
-                    from langparse.workbooks.objects import validate_object_source
-
-                    validate_object_source(workbook_ir.snapshot, source_range)
-                else:
-                    _source_range_is_valid(workbook_ir.snapshot, source_range)
+                _source_range_is_valid(workbook_ir.snapshot, source_range)
             if chunk.metadata["chunk_type"] != "table_rows":
                 continue
             payload = chunk.structured_payload
@@ -186,29 +167,6 @@ class WorkbookStructuralChunker:
         block: WorkbookBlock,
         chunk_index_offset: int,
     ) -> list[Chunk]:
-        if block.kind in {"chart", "image"}:
-            from langparse.workbooks.objects import render_object
-
-            content = render_object(block)
-            metadata = document_metadata(parsed)
-            metadata.update(
-                {
-                    "chunk_type": block.kind,
-                    "chunk_index": chunk_index_offset,
-                    "sheet_name": sheet_name,
-                    "sheet_ordinal": sheet_ordinal,
-                    "source_ranges": [ref.key for ref in block.source_refs],
-                    "object_id": block.block_id,
-                    "oversized": self.length_function(content) > self.max_chunk_size,
-                }
-            )
-            return [
-                Chunk(
-                    content=content,
-                    metadata=metadata,
-                    structured_payload={"object": block.metadata["object"]},
-                )
-            ]
         if block.logical_table is not None:
             return self._chunk_logical_table(
                 parsed,

{langparse-0.1.0 → langparse-0.1.0rc2}/langparse/cli.py

@@ -6,9 +6,7 @@ from collections.abc import Sequence
 from pathlib import Path
 
 from langparse import __version__
-from langparse.chunkers.registry import available_chunkers
 from langparse.errors import classify_exception
-from langparse.progress import ProgressEvent
 from langparse.services.batch_service import BatchParseService
 from langparse.services.benchmark_service import BenchmarkService
 from langparse.services.parse_service import ParseService
@@ -45,27 +43,18 @@ def build_parser():
     parse_cmd.add_argument("--model-source", default=None)
     parse_cmd.add_argument("--auto-install-runtime", action="store_true")
     parse_cmd.add_argument("--runtime-package", default=None)
-    parse_cmd.add_argument("--format", default="markdown", help="markdown, json, or workbook-json")
+    parse_cmd.add_argument("--format", default="markdown")
     parse_cmd.add_argument("--batch", action="store_true")
     parse_cmd.add_argument("--output", default=None)
     parse_cmd.add_argument("--output-dir", default=None)
     parse_cmd.add_argument("--max-workers", type=int, default=None)
     parse_cmd.add_argument("--skip-existing", action="store_true")
     parse_cmd.add_argument("--metrics", action="store_true")
-    parse_cmd.add_argument("--progress", action="store_true", help="write progress to stderr")
     parse_cmd.add_argument(
         "--chunk",
         action="store_true",
         help="semantically chunk the parsed document and include chunks in the output",
     )
-    parse_cmd.add_argument("--chunk-strategy", choices=available_chunkers(), default=None)
-    parse_cmd.add_argument(
-        "--chunk-size",
-        type=int,
-        default=None,
-        help="strategy size budget (lexical tokens for fixed-token, characters otherwise)",
-    )
-    parse_cmd.add_argument("--chunk-overlap", type=int, default=None)
     parse_cmd.add_argument(
         "--chunk-profile",
         choices=["retrieval", "analysis"],
@@ -251,22 +240,6 @@ def _run(args, parser) -> int:
         if value is not None and value is not False
     }
     chunk_kwargs = {"chunk_profile": args.chunk_profile} if args.chunk else {}
-    if any(
-        value is not None for value in (args.chunk_strategy, args.chunk_size, args.chunk_overlap)
-    ):
-        if not args.chunk:
-            parser.error("chunk strategy/size/overlap options require --chunk")
-        chunk_kwargs["chunk_strategy"] = args.chunk_strategy or "semantic"
-        chunk_kwargs["chunk_options"] = {
-            key: value
-            for key, value in {
-                "max_chunk_size": args.chunk_size,
-                "overlap": args.chunk_overlap,
-            }.items()
-            if value is not None
-        }
-    if args.progress:
-        parse_kwargs["progress_callback"] = _print_progress
 
     if args.batch:
         # One implementation regardless of flags. Without --output-dir the run
@@ -309,21 +282,5 @@ def _run(args, parser) -> int:
     return 0
 
 
-def _print_progress(event: ProgressEvent) -> None:
-    counts = ""
-    if event.completed_units is not None:
-        total = event.total_units if event.total_units is not None else "?"
-        counts = f" {event.completed_units}/{total} {event.unit or ''}"
-    percent = f" {event.percent:.0f}%" if event.percent is not None else ""
-    # Escape embedded line breaks/control characters from untrusted filenames/messages.
-    source = repr(event.source) if event.source else "batch"
-    message = f" {event.message!r}" if event.message else ""
-    print(
-        f"langparse: {source} {event.phase} {event.state}{counts}{percent}{message}",
-        file=sys.stderr,
-        flush=True,
-    )
-
-
 if __name__ == "__main__":
     raise SystemExit(main())

{langparse-0.1.0 → langparse-0.1.0rc2}/langparse/engines/pdf/deepdoc_engine.py

@@ -6,7 +6,6 @@ from typing import Any
 from langparse.core.engine import PageResult
 from langparse.engines.pdf.simple import BasePDFEngine
 from langparse.logging import get_logger
-from langparse.progress import ProgressCallback, ProgressReporter
 from langparse.types import ParsedDocumentResult
 
 logger = get_logger(__name__)
@@ -98,15 +97,7 @@ class DeepDocEngine(BasePDFEngine):
             logger.warning("Skipping OCR-applied classification for %s: %s", file_path, exc)
             return {}
 
-    def process_document(
-        self,
-        file_path: Path,
-        *,
-        progress_callback: ProgressCallback | None = None,
-        **kwargs: Any,
-    ) -> ParsedDocumentResult:
-        reporter = ProgressReporter(str(file_path), progress_callback)
-        reporter.emit("preparing", message="Preparing DeepDoc parser")
+    def process_document(self, file_path: Path, **kwargs: Any) -> ParsedDocumentResult:
         try:
             from langparse.engines.pdf.deepdoc.rendering import render_pages
         except ImportError as exc:
@@ -117,19 +108,8 @@ class DeepDocEngine(BasePDFEngine):
         with self._parser_lock:
             if self._parser is None:
                 self._parser = self._build_parser()
-            if progress_callback is None:
-                boxes = self._parser.parse_into_bboxes(str(file_path))
-            else:
-                last_percent = 0.0
-
-                def on_progress(progress, message=""):
-                    nonlocal last_percent
-                    last_percent = max(last_percent, min(100.0, max(0.0, progress * 100)))
-                    reporter.emit("parsing", percent=last_percent, message=message)
-
-                boxes = self._parser.parse_into_bboxes(str(file_path), callback=on_progress)
+            boxes = self._parser.parse_into_bboxes(str(file_path))
 
-        reporter.emit("rendering")
         ocr_pages = self._classify_ocr_pages(file_path)
         pages = render_pages(boxes, ocr_pages=ocr_pages)
         return ParsedDocumentResult(
@@ -146,11 +126,7 @@ class DeepDocEngine(BasePDFEngine):
             },
         )
 
-    def process(
-        self, file_path: Path, *, progress_callback: ProgressCallback | None = None, **kwargs
-    ) -> Iterator[PageResult]:
-        if progress_callback is not None:
-            kwargs["progress_callback"] = progress_callback
+    def process(self, file_path: Path, **kwargs) -> Iterator[PageResult]:
         parsed = self.process_document(file_path, **kwargs)
         for page in parsed.pages:
             yield PageResult(

{langparse-0.1.0 → langparse-0.1.0rc2}/langparse/engines/pdf/mineru.py

@@ -6,7 +6,6 @@ from langparse.core.engine import PageResult
 from langparse.engines.pdf.mineru_client import MinerUClient
 from langparse.engines.pdf.mineru_service import MinerUServiceManager
 from langparse.engines.pdf.simple import BasePDFEngine
-from langparse.progress import ProgressCallback, ProgressReporter
 from langparse.types import ParsedDocumentResult, ParsedElement, ParsedPageResult
 
 _SENSITIVE_OPTION_KEYS = frozenset(
@@ -152,35 +151,16 @@ class MinerUEngine(BasePDFEngine):
     def _create_service_manager(self) -> MinerUServiceManager:
         return MinerUServiceManager(**self._build_service_config())
 
-    def _run_mineru(
-        self,
-        file_path: Path,
-        runtime_config: dict[str, Any],
-        progress_callback: ProgressCallback | None = None,
-    ) -> list[dict[str, Any]]:
-        reporter = ProgressReporter(str(file_path), progress_callback)
+    def _run_mineru(self, file_path: Path, runtime_config: dict[str, Any]) -> list[dict[str, Any]]:
         manager = self._create_service_manager()
         with manager.running_service() as base_url:
             client = self._create_client(base_url)
-            reporter.emit("parsing", message="Waiting for MinerU response")
             return client.parse_file(file_path, runtime_config)
 
-    def process_document(
-        self,
-        file_path: Path,
-        *,
-        progress_callback: ProgressCallback | None = None,
-        **kwargs: Any,
-    ) -> ParsedDocumentResult:
-        reporter = ProgressReporter(str(file_path), progress_callback)
-        reporter.emit("preparing", message="Preparing MinerU service")
+    def process_document(self, file_path: Path, **kwargs: Any) -> ParsedDocumentResult:
         self._ensure_runtime()
         runtime_config = self._build_runtime_config(**kwargs)
-        if progress_callback is None:
-            raw_pages = self._run_mineru(file_path, runtime_config)
-        else:
-            raw_pages = self._run_mineru(file_path, runtime_config, progress_callback)
-        reporter.emit("rendering")
+        raw_pages = self._run_mineru(file_path, runtime_config)
         pages = [
             ParsedPageResult(
                 page_number=item["page_number"],
@@ -241,11 +221,7 @@ class MinerUEngine(BasePDFEngine):
             },
         )
 
-    def process(
-        self, file_path: Path, *, progress_callback: ProgressCallback | None = None, **kwargs
-    ) -> Iterator[PageResult]:
-        if progress_callback is not None:
-            kwargs["progress_callback"] = progress_callback
+    def process(self, file_path: Path, **kwargs) -> Iterator[PageResult]:
         parsed = self.process_document(file_path, **kwargs)
         for page in parsed.pages:
             yield PageResult(

{langparse-0.1.0 → langparse-0.1.0rc2}/langparse/engines/pdf/simple.py

@@ -10,7 +10,6 @@ from langparse.engines.pdf.ocr import (
     needs_ocr,
     ocr_page_text,
 )
-from langparse.progress import ProgressCallback, ProgressReporter
 
 
 class BasePDFEngine(BaseEngine):
@@ -50,10 +49,7 @@ class SimplePDFEngine(BasePDFEngine):
         # guarantee -- correctness over throughput on an already slow path.
         self._ocr_lock = threading.Lock()
 
-    def process(
-        self, file_path: Path, *, progress_callback: ProgressCallback | None = None, **kwargs
-    ) -> Iterator[PageResult]:
-        reporter = ProgressReporter(str(file_path), progress_callback)
+    def process(self, file_path: Path, **kwargs) -> Iterator[PageResult]:
         try:
             import pdfplumber
         except ImportError:
@@ -76,9 +72,6 @@ class SimplePDFEngine(BasePDFEngine):
                 if table_markdown:
                     markdown_content = "\n\n".join([text, "\n".join(table_markdown)]).strip()
 
-                reporter.emit(
-                    "parsing", completed_units=i + 1, total_units=len(pdf.pages), unit="pages"
-                )
                 yield PageResult(
                     page_number=i + 1,
                     markdown_content=markdown_content,

{langparse-0.1.0 → langparse-0.1.0rc2}/langparse/errors.py

@@ -1,10 +1,7 @@
 from __future__ import annotations
 
-import re
 from dataclasses import dataclass
 from enum import Enum
-from subprocess import TimeoutExpired
-from urllib.error import URLError
 
 
 class ErrorType(str, Enum):
@@ -45,7 +42,7 @@ def classify_exception(exc: BaseException) -> ClassifiedError:
         return ClassifiedError(ErrorType.ENGINE_UNAVAILABLE, message)
     if "unable to start local mineru-api" in lowered:
         return ClassifiedError(ErrorType.ENGINE_UNAVAILABLE, message)
-    if _is_timeout(exc) or re.search(r"\btimed out\b", lowered):
+    if "timed out" in lowered or "timeout" in lowered:
         return ClassifiedError(ErrorType.ENGINE_TIMEOUT, message)
     if "ocr" in lowered and "unavailable" in lowered:
         return ClassifiedError(ErrorType.OCR_UNAVAILABLE, message)
@@ -53,18 +50,3 @@ def classify_exception(exc: BaseException) -> ClassifiedError:
         return ClassifiedError(ErrorType.TABLE_EXTRACTION_FAILED, message)
 
     return ClassifiedError(ErrorType.PARSE_FAILED, message)
-
-
-def _is_timeout(exc: BaseException) -> bool:
-    """Recognize standard timeout types through explicit backend wrappers."""
-    seen: set[int] = set()
-    current: BaseException | None = exc
-    while current is not None and id(current) not in seen:
-        seen.add(id(current))
-        if isinstance(current, (TimeoutError, TimeoutExpired)):
-            return True
-        if isinstance(current, URLError) and isinstance(current.reason, BaseException):
-            current = current.reason
-        else:
-            current = current.__cause__
-    return False

{langparse-0.1.0 → langparse-0.1.0rc2}/langparse/parsers/excel_parser.py

@@ -3,7 +3,6 @@ from pathlib import Path
 
 from langparse.core.parser import BaseParser
 from langparse.parsers.sniff import looks_like_ole_binary, looks_like_zip_ooxml
-from langparse.progress import ProgressCallback, ProgressReporter
 from langparse.types import ParsedDocumentResult, ParsedElement, ParseDiagnostics, ParsedPageResult
 from langparse.workbooks.modeling import (
     RequiredWorkbookDisambiguationError,
@@ -65,14 +64,11 @@ class ExcelParser(BaseParser):
                 WorkbookDisambiguation.off() if disambiguation is None else disambiguation
             )
 
-    def parse_result(
-        self, file_path: str | Path, *, progress_callback: ProgressCallback | None = None, **kwargs
-    ) -> ParsedDocumentResult:
+    def parse_result(self, file_path: str | Path, **kwargs) -> ParsedDocumentResult:
         path = self._resolve_existing_path(file_path)
-        reporter = ProgressReporter(path, progress_callback)
 
         if looks_like_zip_ooxml(path):
-            return self._parse_ooxml(path, reporter)
+            return self._parse_ooxml(path)
 
         try:
             import pandas as pd
@@ -86,7 +82,6 @@ class ExcelParser(BaseParser):
         # otherwise be handed to the wrong pandas reader. Content decides:
         # a real workbook is either a ZIP-OOXML or legacy-OLE container;
         # anything else is read as delimited text regardless of its label.
-        reporter.emit("extracting")
         is_legacy_workbook = looks_like_ole_binary(path)
         if is_legacy_workbook:
             sheets = pd.read_excel(path, sheet_name=None)
@@ -110,7 +105,6 @@ class ExcelParser(BaseParser):
                 )
             }
 
-        reporter.emit("rendering")
         pages = [
             self._page_for_sheet(index + 1, sheet_name, frame)
             for index, (sheet_name, frame) in enumerate(sheets.items())
@@ -138,7 +132,7 @@ class ExcelParser(BaseParser):
             ),
         )
 
-    def _parse_ooxml(self, path: Path, reporter: ProgressReporter) -> ParsedDocumentResult:
+    def _parse_ooxml(self, path: Path) -> ParsedDocumentResult:
         try:
             from langparse.workbooks.adapters import OOXMLWorkbookAdapter
             from langparse.workbooks.assembly import assemble_baseline, assemble_workbook
@@ -152,9 +146,7 @@ class ExcelParser(BaseParser):
                 "Install with `pip install langparse[excel]`."
             ) from None
 
-        reporter.emit("extracting")
         snapshot = OOXMLWorkbookAdapter().snapshot(path)
-        reporter.emit("assembling")
         try:
             structure, diagnostics = assemble_workbook(
                 snapshot,
@@ -168,10 +160,6 @@ class ExcelParser(BaseParser):
             diagnostics.warnings.append(
                 f"Semantic workbook assembly failed; retained raw-grid fallback: {type(exc).__name__}"
             )
-        from langparse.workbooks.objects import attach_objects
-
-        attach_objects(snapshot, structure, diagnostics)
-        reporter.emit("rendering")
         pages = compatibility_pages(snapshot, structure)
         markdown = render_workbook_markdown(snapshot, structure)
         return ParsedDocumentResult(