man1lab 1.2.2__tar.gz

man1lab-1.2.2/.env.example

@@ -0,0 +1,7 @@
+# DeepSeek (OpenAI-compatible) — copy to .env and fill in your key
+OPENAI_API_KEY=sk-your-key-here
+OPENAI_BASE_URL=https://api.deepseek.com
+OPENAI_MODEL=deepseek-v4-pro
+
+# Paper parsing backend: docling (default) or pymupdf (legacy fallback)
+PARSER_BACKEND=docling

man1lab-1.2.2/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 maniac1um
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

man1lab-1.2.2/MANIFEST.in

@@ -0,0 +1,13 @@
+include LICENSE
+include README.md
+include .env.example
+recursive-include conf *
+recursive-include prompts *
+prune workspace/tasks
+prune outputs
+prune logs
+prune mlruns
+prune .cache
+global-exclude __pycache__
+global-exclude *.py[cod]
+global-exclude .DS_Store

man1lab-1.2.2/PKG-INFO

@@ -0,0 +1,240 @@
+Metadata-Version: 2.4
+Name: man1lab
+Version: 1.2.2
+Summary: Autonomous research paper reproduction platform
+Author: maniac1um
+License: MIT
+Project-URL: Homepage, https://github.com/maniac1um/Man1Lab
+Project-URL: Repository, https://github.com/maniac1um/Man1Lab
+Project-URL: Documentation, https://github.com/maniac1um/Man1Lab/blob/main/docs/GETTING_STARTED.md
+Keywords: research,reproduction,machine-learning,agents,pipeline
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Science/Research
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: pydantic>=2.0
+Requires-Dist: pymupdf>=1.24.0
+Requires-Dist: docling>=2.0.0
+Requires-Dist: openai>=1.0.0
+Requires-Dist: anthropic>=0.40.0
+Requires-Dist: python-dotenv>=1.0.0
+Requires-Dist: hydra-core>=1.3.0
+Requires-Dist: mlflow>=2.0.0
+Requires-Dist: httpx>=0.28.0
+Requires-Dist: typer>=0.15.0
+Provides-Extra: dev
+Requires-Dist: pytest>=8.0; extra == "dev"
+Requires-Dist: build>=1.0; extra == "dev"
+Dynamic: license-file
+
+# Man1Lab
+
+**Engineering-first autonomous research paper reproduction platform.**
+
+Man1Lab automates the engineering workflow behind AI paper reproduction — from reading a PDF to discovering official code and checkpoints, planning how to run the work, and driving implementation through a structured pipeline. One command. One platform.
+
+[![PyPI version](https://img.shields.io/pypi/v/man1lab)](https://pypi.org/project/man1lab/)
+[![Python Version](https://img.shields.io/pypi/pyversions/man1lab)](https://pypi.org/project/man1lab/)
+[![Latest Release](https://img.shields.io/github/v/release/maniac1um/Man1Lab)](https://github.com/maniac1um/Man1Lab/releases)
+[![Tests](https://img.shields.io/badge/tests-614%20passing-brightgreen)](docs/CURRENT_STATUS.md)
+[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](LICENSE)
+[![Documentation](https://img.shields.io/badge/docs-Getting%20Started-blue)](docs/GETTING_STARTED.md)
+
+---
+
+## Why Man1Lab?
+
+Reproducing a modern AI paper is an engineering project, not a reading exercise.
+
+Researchers and engineers routinely spend days on work that follows the same pattern:
+
+| Pain point | What you do today |
+|------------|-------------------|
+| Understand the paper | Read, highlight, and manually extract methods and requirements |
+| Find official resources | Search GitHub, Hugging Face, and project pages by hand |
+| Locate checkpoints and data | Track down weights, configs, and dataset links across sites |
+| Prepare the environment | Create repos, install dependencies, wire configs manually |
+| Bridge paper → code | Map sections of the paper to modules, scripts, and training steps |
+| Connect everything | Stitch analysis, resources, and execution into one coherent plan |
+
+**Man1Lab automates this engineering workflow.**
+
+Give it a paper. It produces structured analysis, discovers resources, commits an execution strategy, and runs the reproduction pipeline — so you start from engineering decisions, not from scratch.
+
+---
+
+## Core Workflow
+
+```text
+Paper
+  ↓
+Analysis          — Extract reproduction requirements from the PDF
+  ↓
+Discovery         — Find official repositories, checkpoints, and datasets
+  ↓
+Execution Planning — Commit how the work should be engineered
+  ↓
+Execution         — Plan tasks, generate code, run training, verify, report
+```
+
+Each stage produces a structured artifact that feeds the next. The full pipeline is available through the CLI and Python SDK.
+
+---
+
+## Key Features
+
+| Capability | Description | Status |
+|------------|-------------|--------|
+| Paper Analysis | Structured extraction from research PDFs | ✅ |
+| Research Resource Discovery | Evidence-backed search for repos, models, and data | ✅ |
+| Execution Planning | Engineering strategy before implementation | ✅ |
+| CLI | `man1lab` — init, reproduce, model management, and more | ✅ |
+| Python SDK | `from man1lab import Man1Lab` | ✅ |
+| Multi-model Support | Switch LLM providers without editing config files | ✅ |
+| Model Registry | Named profiles, active model, portable export/import | ✅ |
+| OpenAI | GPT-family models | ✅ |
+| DeepSeek | DeepSeek API-compatible models | ✅ |
+| Anthropic | Claude models | ✅ |
+| Package Distribution | `pip install man1lab` | ✅ |
+| Platform Facade | Single entry point for CLI, SDK, and future interfaces | ✅ |
+
+---
+
+## Current Scope
+
+Man1Lab targets **engineering-oriented software reproduction** of AI research papers.
+
+| Domain | Supported as software reproduction |
+|--------|-------------------------------------|
+| Computer Vision | ✅ |
+| Embodied AI (software stack) | ✅ |
+| LLM Systems | ✅ |
+| Agent Frameworks | ✅ |
+| Reinforcement Learning | ✅ |
+
+**What Man1Lab does:** automate analysis, resource discovery, execution planning, and the software reproduction pipeline.
+
+**What is outside current scope:** physical robot deployment, hardware setup, calibration, and real-world deployment. Those remain manual engineering work. Man1Lab focuses on the software path from paper to runnable reproduction.
+
+---
+
+## Quick Start
+
+```bash
+pip install man1lab
+man1lab init
+man1lab doctor
+man1lab reproduce paper.pdf
+```
+
+During `init`, you can configure your first LLM provider interactively. Full installation and configuration: [Getting Started](docs/GETTING_STARTED.md).
+
+---
+
+## Example Workflow
+
+```bash
+man1lab reproduce OpenVLA.pdf
+```
+
+```text
+Analysis completed
+  ↓
+Official repository discovered
+  ↓
+Checkpoint and dataset candidates ranked
+  ↓
+Execution strategy generated
+  ↓
+Engineering tasks planned → repository built → training run → report
+```
+
+The platform handles the chain from paper to execution. You review outputs and iterate on results.
+
+---
+
+## Project Architecture
+
+```text
+Platform
+  ↓
+Analysis → Discovery → Execution Planning → Execution
+```
+
+Man1Lab is organized as a single platform with a clear stage-by-stage pipeline. Interfaces (CLI, SDK) delegate to one composition root — no direct coupling to internal agents from user code.
+
+Details: [Architecture](docs/architecture/ARCHITECTURE.md)
+
+---
+
+## Roadmap
+
+| Version | Focus | Status |
+|---------|-------|--------|
+| **v1.2.x** | Platform foundation, Execution Planning, CLI, SDK, multi-model support | **Completed** |
+| **v1.3** | Repository Understanding — semantic mapping of discovered code to paper modules | Planned |
+| **v1.4** | Repository Adaptation — align discovered repos with paper requirements | Planned |
+| **v1.5** | Knowledge Memory — cross-run reproduction knowledge | Planned |
+
+Live status: [Current Status](docs/CURRENT_STATUS.md) · Full roadmap: [ROADMAP.md](ROADMAP.md)
+
+---
+
+## Documentation
+
+| Document | Description |
+|----------|-------------|
+| [Getting Started](docs/GETTING_STARTED.md) | Install, configure models, run your first reproduction |
+| [Architecture](docs/architecture/ARCHITECTURE.md) | Platform design and canonical artifacts |
+| [Current Status](docs/CURRENT_STATUS.md) | Capabilities, tests, and known limitations |
+| [Release Notes](docs/releases/v1.2.2.md) | v1.2.2 — LLM platform and first-run experience |
+| [Contributing](CONTRIBUTING.md) | Development setup and contribution guidelines |
+| [Security](SECURITY.md) | Vulnerability reporting |
+| [Support](SUPPORT.md) | Questions, bugs, and feature requests |
+
+---
+
+## Community
+
+| Channel | Use for |
+|---------|---------|
+| [GitHub Issues](https://github.com/maniac1um/Man1Lab/issues) | Bug reports and feature requests |
+| [GitHub Discussions](https://github.com/maniac1um/Man1Lab/discussions) | Questions and ideas |
+| [Pull Requests](https://github.com/maniac1um/Man1Lab/pulls) | Code contributions ([guidelines](CONTRIBUTING.md)) |
+| [Security Advisories](https://github.com/maniac1um/Man1Lab/security/advisories/new) | Private vulnerability reports only |
+
+Man1Lab is an active research prototype for academic demonstration. See [Contributing](CONTRIBUTING.md) before opening large changes.
+
+---
+
+## Citation
+
+```bibtex
+@software{man1lab_2026,
+  author       = {maniac1um},
+  title        = {Man1Lab: An Autonomous Research Paper Reproduction Platform},
+  year         = {2026},
+  version      = {1.2.2},
+  url          = {https://github.com/maniac1um/Man1Lab},
+  note         = {Engineering-first autonomous research paper reproduction platform.}
+}
+```
+
+Also see [CITATION.cff](CITATION.cff).
+
+---
+
+## License
+
+Man1Lab is released under the MIT License. See the `LICENSE` file for details.
+
+---
+
+**Maintainer:** [maniac1um](https://github.com/maniac1um)

man1lab-1.2.2/README.md

@@ -0,0 +1,204 @@
+# Man1Lab
+
+**Engineering-first autonomous research paper reproduction platform.**
+
+Man1Lab automates the engineering workflow behind AI paper reproduction — from reading a PDF to discovering official code and checkpoints, planning how to run the work, and driving implementation through a structured pipeline. One command. One platform.
+
+[![PyPI version](https://img.shields.io/pypi/v/man1lab)](https://pypi.org/project/man1lab/)
+[![Python Version](https://img.shields.io/pypi/pyversions/man1lab)](https://pypi.org/project/man1lab/)
+[![Latest Release](https://img.shields.io/github/v/release/maniac1um/Man1Lab)](https://github.com/maniac1um/Man1Lab/releases)
+[![Tests](https://img.shields.io/badge/tests-614%20passing-brightgreen)](docs/CURRENT_STATUS.md)
+[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](LICENSE)
+[![Documentation](https://img.shields.io/badge/docs-Getting%20Started-blue)](docs/GETTING_STARTED.md)
+
+---
+
+## Why Man1Lab?
+
+Reproducing a modern AI paper is an engineering project, not a reading exercise.
+
+Researchers and engineers routinely spend days on work that follows the same pattern:
+
+| Pain point | What you do today |
+|------------|-------------------|
+| Understand the paper | Read, highlight, and manually extract methods and requirements |
+| Find official resources | Search GitHub, Hugging Face, and project pages by hand |
+| Locate checkpoints and data | Track down weights, configs, and dataset links across sites |
+| Prepare the environment | Create repos, install dependencies, wire configs manually |
+| Bridge paper → code | Map sections of the paper to modules, scripts, and training steps |
+| Connect everything | Stitch analysis, resources, and execution into one coherent plan |
+
+**Man1Lab automates this engineering workflow.**
+
+Give it a paper. It produces structured analysis, discovers resources, commits an execution strategy, and runs the reproduction pipeline — so you start from engineering decisions, not from scratch.
+
+---
+
+## Core Workflow
+
+```text
+Paper
+  ↓
+Analysis          — Extract reproduction requirements from the PDF
+  ↓
+Discovery         — Find official repositories, checkpoints, and datasets
+  ↓
+Execution Planning — Commit how the work should be engineered
+  ↓
+Execution         — Plan tasks, generate code, run training, verify, report
+```
+
+Each stage produces a structured artifact that feeds the next. The full pipeline is available through the CLI and Python SDK.
+
+---
+
+## Key Features
+
+| Capability | Description | Status |
+|------------|-------------|--------|
+| Paper Analysis | Structured extraction from research PDFs | ✅ |
+| Research Resource Discovery | Evidence-backed search for repos, models, and data | ✅ |
+| Execution Planning | Engineering strategy before implementation | ✅ |
+| CLI | `man1lab` — init, reproduce, model management, and more | ✅ |
+| Python SDK | `from man1lab import Man1Lab` | ✅ |
+| Multi-model Support | Switch LLM providers without editing config files | ✅ |
+| Model Registry | Named profiles, active model, portable export/import | ✅ |
+| OpenAI | GPT-family models | ✅ |
+| DeepSeek | DeepSeek API-compatible models | ✅ |
+| Anthropic | Claude models | ✅ |
+| Package Distribution | `pip install man1lab` | ✅ |
+| Platform Facade | Single entry point for CLI, SDK, and future interfaces | ✅ |
+
+---
+
+## Current Scope
+
+Man1Lab targets **engineering-oriented software reproduction** of AI research papers.
+
+| Domain | Supported as software reproduction |
+|--------|-------------------------------------|
+| Computer Vision | ✅ |
+| Embodied AI (software stack) | ✅ |
+| LLM Systems | ✅ |
+| Agent Frameworks | ✅ |
+| Reinforcement Learning | ✅ |
+
+**What Man1Lab does:** automate analysis, resource discovery, execution planning, and the software reproduction pipeline.
+
+**What is outside current scope:** physical robot deployment, hardware setup, calibration, and real-world deployment. Those remain manual engineering work. Man1Lab focuses on the software path from paper to runnable reproduction.
+
+---
+
+## Quick Start
+
+```bash
+pip install man1lab
+man1lab init
+man1lab doctor
+man1lab reproduce paper.pdf
+```
+
+During `init`, you can configure your first LLM provider interactively. Full installation and configuration: [Getting Started](docs/GETTING_STARTED.md).
+
+---
+
+## Example Workflow
+
+```bash
+man1lab reproduce OpenVLA.pdf
+```
+
+```text
+Analysis completed
+  ↓
+Official repository discovered
+  ↓
+Checkpoint and dataset candidates ranked
+  ↓
+Execution strategy generated
+  ↓
+Engineering tasks planned → repository built → training run → report
+```
+
+The platform handles the chain from paper to execution. You review outputs and iterate on results.
+
+---
+
+## Project Architecture
+
+```text
+Platform
+  ↓
+Analysis → Discovery → Execution Planning → Execution
+```
+
+Man1Lab is organized as a single platform with a clear stage-by-stage pipeline. Interfaces (CLI, SDK) delegate to one composition root — no direct coupling to internal agents from user code.
+
+Details: [Architecture](docs/architecture/ARCHITECTURE.md)
+
+---
+
+## Roadmap
+
+| Version | Focus | Status |
+|---------|-------|--------|
+| **v1.2.x** | Platform foundation, Execution Planning, CLI, SDK, multi-model support | **Completed** |
+| **v1.3** | Repository Understanding — semantic mapping of discovered code to paper modules | Planned |
+| **v1.4** | Repository Adaptation — align discovered repos with paper requirements | Planned |
+| **v1.5** | Knowledge Memory — cross-run reproduction knowledge | Planned |
+
+Live status: [Current Status](docs/CURRENT_STATUS.md) · Full roadmap: [ROADMAP.md](ROADMAP.md)
+
+---
+
+## Documentation
+
+| Document | Description |
+|----------|-------------|
+| [Getting Started](docs/GETTING_STARTED.md) | Install, configure models, run your first reproduction |
+| [Architecture](docs/architecture/ARCHITECTURE.md) | Platform design and canonical artifacts |
+| [Current Status](docs/CURRENT_STATUS.md) | Capabilities, tests, and known limitations |
+| [Release Notes](docs/releases/v1.2.2.md) | v1.2.2 — LLM platform and first-run experience |
+| [Contributing](CONTRIBUTING.md) | Development setup and contribution guidelines |
+| [Security](SECURITY.md) | Vulnerability reporting |
+| [Support](SUPPORT.md) | Questions, bugs, and feature requests |
+
+---
+
+## Community
+
+| Channel | Use for |
+|---------|---------|
+| [GitHub Issues](https://github.com/maniac1um/Man1Lab/issues) | Bug reports and feature requests |
+| [GitHub Discussions](https://github.com/maniac1um/Man1Lab/discussions) | Questions and ideas |
+| [Pull Requests](https://github.com/maniac1um/Man1Lab/pulls) | Code contributions ([guidelines](CONTRIBUTING.md)) |
+| [Security Advisories](https://github.com/maniac1um/Man1Lab/security/advisories/new) | Private vulnerability reports only |
+
+Man1Lab is an active research prototype for academic demonstration. See [Contributing](CONTRIBUTING.md) before opening large changes.
+
+---
+
+## Citation
+
+```bibtex
+@software{man1lab_2026,
+  author       = {maniac1um},
+  title        = {Man1Lab: An Autonomous Research Paper Reproduction Platform},
+  year         = {2026},
+  version      = {1.2.2},
+  url          = {https://github.com/maniac1um/Man1Lab},
+  note         = {Engineering-first autonomous research paper reproduction platform.}
+}
+```
+
+Also see [CITATION.cff](CITATION.cff).
+
+---
+
+## License
+
+Man1Lab is released under the MIT License. See the `LICENSE` file for details.
+
+---
+
+**Maintainer:** [maniac1um](https://github.com/maniac1um)

man1lab-1.2.2/adapters/__init__.py

@@ -0,0 +1,4 @@
+from adapters.config_parser_settings import ConfigParserSettingsProvider
+from adapters.factory import build_document_parser
+
+__all__ = ["ConfigParserSettingsProvider", "build_document_parser"]

man1lab-1.2.2/adapters/config_parser_settings.py

@@ -0,0 +1,18 @@
+from configuration.provider import SettingsProvider, get_settings_provider
+from ports.output_format import OutputFormat
+from ports.parser_settings import ParserSettings, ParserSettingsProvider
+
+
+class ConfigParserSettingsProvider:
+    """Read parser settings from the application settings provider."""
+
+    def __init__(self, settings_provider: SettingsProvider | None = None) -> None:
+        self._settings_provider = settings_provider or get_settings_provider()
+
+    def get_parser_settings(self) -> ParserSettings:
+        settings = self._settings_provider.get_settings()
+        return ParserSettings(
+            backend=settings.parser.backend.strip().lower(),
+            output_format=OutputFormat.MARKDOWN,
+            max_paper_text_chars=settings.parser.max_paper_text_chars,
+        )

man1lab-1.2.2/adapters/docling_parser.py

@@ -0,0 +1,88 @@
+import logging
+import time
+from pathlib import Path
+
+from docling.backend.pypdfium2_backend import PyPdfiumDocumentBackend
+from docling.datamodel.base_models import InputFormat
+from docling.datamodel.pipeline_options import PdfPipelineOptions
+from docling.document_converter import DocumentConverter, PdfFormatOption
+
+from adapters.output_export import export_docling_document
+from ports.output_format import OutputFormat
+from ports.parsed_document import ParsedDocument
+from services.exceptions import (
+    PDFEmptyError,
+    PDFEncryptedError,
+    PDFExtractionError,
+    PDFNotFoundError,
+)
+
+logger = logging.getLogger(__name__)
+
+
+class DoclingParser:
+    """Document parser backed by Docling structured document export."""
+
+    def __init__(
+        self,
+        converter: DocumentConverter | None = None,
+        *,
+        output_format: OutputFormat = OutputFormat.MARKDOWN,
+        max_paper_text_chars: int = 80_000,
+    ) -> None:
+        self._converter = converter or _build_converter()
+        self._output_format = output_format
+        self._max_paper_text_chars = max_paper_text_chars
+
+    def parse(self, paper_path: Path) -> ParsedDocument:
+        if not paper_path.exists():
+            raise PDFNotFoundError(f"PDF not found: {paper_path}")
+
+        start = time.perf_counter()
+        try:
+            result = self._converter.convert(str(paper_path))
+        except Exception as exc:
+            message = str(exc).lower()
+            if "encrypt" in message or "password" in message:
+                raise PDFEncryptedError(
+                    f"PDF is encrypted and cannot be read: {paper_path}"
+                ) from exc
+            raise PDFExtractionError(
+                f"Failed to parse PDF with Docling: {paper_path}"
+            ) from exc
+
+        markdown = export_docling_document(
+            result.document,
+            self._output_format,
+        )
+        if not markdown:
+            raise PDFEmptyError(f"PDF contains no extractable text: {paper_path}")
+
+        if self._max_paper_text_chars and len(markdown) > self._max_paper_text_chars:
+            markdown = markdown[: self._max_paper_text_chars]
+
+        duration = time.perf_counter() - start
+        logger.info(
+            "Docling parse complete: chars=%d duration=%.3fs file=%s status=%s format=%s",
+            len(markdown),
+            duration,
+            paper_path.name,
+            getattr(result, "status", "unknown"),
+            self._output_format.value,
+        )
+        return ParsedDocument(markdown=markdown)
+
+
+def _build_converter() -> DocumentConverter:
+    pipeline_options = PdfPipelineOptions(
+        do_ocr=False,
+        do_table_structure=True,
+    )
+    return DocumentConverter(
+        format_options={
+            InputFormat.PDF: PdfFormatOption(
+                pipeline_options=pipeline_options,
+                backend=PyPdfiumDocumentBackend,
+            )
+        }
+    )

man1lab-1.2.2/adapters/factory.py

@@ -0,0 +1,31 @@
+from adapters.config_parser_settings import ConfigParserSettingsProvider
+from adapters.docling_parser import DoclingParser
+from adapters.pymupdf_parser import PyMuPDFParser
+from ports.document_parser import DocumentParser
+from ports.parser_settings import ParserSettings, ParserSettingsProvider
+
+_SUPPORTED_BACKENDS = frozenset({"docling", "pymupdf"})
+
+
+def build_document_parser(
+    settings: ParserSettings | None = None,
+    settings_provider: ParserSettingsProvider | None = None,
+) -> DocumentParser:
+    """Construct a document parser from explicit or provider-supplied settings."""
+    if settings is None:
+        provider = settings_provider or ConfigParserSettingsProvider()
+        settings = provider.get_parser_settings()
+
+    backend = settings.backend.strip().lower()
+    if backend not in _SUPPORTED_BACKENDS:
+        supported = ", ".join(sorted(_SUPPORTED_BACKENDS))
+        raise ValueError(
+            f"Unsupported parser backend={settings.backend!r}. "
+            f"Expected one of: {supported}"
+        )
+    if backend == "pymupdf":
+        return PyMuPDFParser()
+    return DoclingParser(
+        output_format=settings.output_format,
+        max_paper_text_chars=settings.max_paper_text_chars,
+    )

man1lab-1.2.2/adapters/output_export.py

@@ -0,0 +1,21 @@
+from typing import Any
+
+from ports.output_format import OutputFormat
+
+
+def export_docling_document(
+    document: Any,
+    output_format: OutputFormat,
+) -> str:
+    """Serialize a Docling document using the configured output format."""
+    if output_format is OutputFormat.MARKDOWN:
+        return document.export_to_markdown().strip()
+    if output_format is OutputFormat.JSON:
+        raise NotImplementedError(
+            "OutputFormat.JSON is reserved for future Docling export support."
+        )
+    if output_format is OutputFormat.DOCTAGS:
+        raise NotImplementedError(
+            "OutputFormat.DOCTAGS is reserved for future Docling export support."
+        )
+    raise ValueError(f"Unsupported output format: {output_format!r}")

man1lab-1.2.2/adapters/pymupdf_parser.py

@@ -0,0 +1,15 @@
+from pathlib import Path
+
+from ports.parsed_document import ParsedDocument
+from services.pdf_service import PDFService
+
+
+class PyMuPDFParser:
+    """Document parser backed by PyMuPDF plain-text extraction (legacy fallback)."""
+
+    def __init__(self, pdf_service: PDFService | None = None) -> None:
+        self._pdf_service = pdf_service or PDFService()
+
+    def parse(self, paper_path: Path) -> ParsedDocument:
+        text = self._pdf_service.extract(paper_path)
+        return ParsedDocument(markdown=text)

man1lab-1.2.2/agents/__init__.py

@@ -0,0 +1,8 @@
+from agents.coder import Coder
+from agents.planner import Planner
+from agents.reader import Reader
+from agents.reporter import Reporter
+from agents.reviewer import Reviewer
+from agents.runner import Runner
+
+__all__ = ["Coder", "Planner", "Reader", "Reporter", "Reviewer", "Runner"]