aimd-cli 0.9.2__py3-none-any.whl

aimd/infrastructure/markitdown_processor.py

@@ -0,0 +1,103 @@
+"""MarkItDown-backed local file conversion."""
+
+import asyncio
+from functools import partial
+from pathlib import Path
+
+from markitdown import MarkItDown
+
+from ..const import BOOK_EXTENSIONS, MARKITDOWN_FILE_EXTENSIONS
+from ..errors import InputNotFoundError, UnsupportedInputError
+from ..types import TextContext
+from .documents.chunking import (
+    combine_sections_for_processing,
+    split_markdown_by_headers,
+)
+from .documents.title_extractor import extract_title_from_content
+
+
+def is_supported_file(file_path: str | Path) -> bool:
+    """Return whether a local file extension should be offered to MarkItDown."""
+    if isinstance(file_path, str) and file_path.startswith(("http://", "https://")):
+        return False
+    return Path(file_path).suffix.lower() in MARKITDOWN_FILE_EXTENSIONS
+
+
+def _text_context_from_markdown(
+    markdown: str,
+    fallback_title: str,
+    title: str | None,
+    max_chunk_size: int,
+) -> TextContext:
+    """Convert MarkItDown markdown output into aimd's TextContext shape."""
+    resolved_title = title or extract_title_from_content(markdown, fallback_title)
+    stripped = markdown.strip()
+    if len(stripped) <= max_chunk_size:
+        return TextContext(
+            title=resolved_title,
+            chunk_list=[stripped] if stripped else [],
+            split_header_level=None,
+        )
+
+    sections, header_level = split_markdown_by_headers(
+        stripped,
+        max_chunk_size=max_chunk_size,
+    )
+    section_data = [
+        (section_title, section_content.strip())
+        for section_title, section_content in sections
+        if section_content.strip()
+    ]
+    chunks = combine_sections_for_processing(section_data, max_chunk_size)
+    return TextContext(
+        title=resolved_title,
+        chunk_list=chunks,
+        split_header_level=header_level,
+    )
+
+
+async def convert_file_with_markitdown(
+    file_path: str | Path,
+    transcribe_engine: str = "auto",
+    language: str | None = None,
+    model: str | None = None,
+    temp_dir: Path | None = None,
+    *,
+    max_chunk_size: int = 40000,
+) -> tuple[TextContext, Path | None]:
+    """Convert a local file through MarkItDown and installed aimd plugins."""
+    input_path = Path(file_path)
+    if not input_path.exists():
+        raise InputNotFoundError(f"Input file not found: {input_path}")
+    if not input_path.is_file():
+        raise UnsupportedInputError(f"Path is not a file: {input_path}")
+
+    suffix = input_path.suffix.lower()
+    output_dir = (
+        input_path.parent / input_path.stem if suffix in BOOK_EXTENSIONS else None
+    )
+
+    md = MarkItDown(enable_plugins=True)
+    loop = asyncio.get_running_loop()
+    result = await loop.run_in_executor(
+        None,
+        partial(
+            md.convert,
+            input_path,
+            transcribe_engine=transcribe_engine,
+            language=language,
+            model=model,
+            temp_dir=temp_dir,
+            output_dir=output_dir,
+        ),
+    )
+    markdown = result.markdown
+    return (
+        _text_context_from_markdown(
+            markdown,
+            fallback_title=input_path.stem,
+            title=result.title,
+            max_chunk_size=max_chunk_size,
+        ),
+        output_dir,
+    )

aimd/infrastructure/media_processor.py

@@ -0,0 +1,51 @@
+"""aimd wrappers around the aimd-media package."""
+
+from pathlib import Path
+
+from aimd_media.errors import ProcessingFailedError as MediaProcessingFailedError
+from aimd_media.errors import UnsupportedInputError as MediaUnsupportedInputError
+from aimd_media.url import get_text_from_url
+
+from ..errors import ProcessingFailedError, UnsupportedInputError
+from ..types import TextContext
+from .markitdown_processor import _text_context_from_markdown
+
+
+async def get_text_context_from_media_url(
+    url: str,
+    transcribe_engine: str = "auto",
+    language: str | None = None,
+    model: str | None = None,
+    save_original_path: Path | None = None,
+    cookies_file: str | None = None,
+    cookies_from_browser: str | None = None,
+    temp_dir: Path | None = None,
+    raw_transcript: bool = False,
+) -> tuple[TextContext, str]:
+    """Extract a media URL through aimd-media and wrap it as TextContext."""
+    try:
+        result = await get_text_from_url(
+            url=url,
+            transcribe_engine=transcribe_engine,
+            language=language,
+            model=model,
+            save_original_path=save_original_path,
+            cookies_file=cookies_file,
+            cookies_from_browser=cookies_from_browser,
+            temp_dir=temp_dir,
+            raw_transcript=raw_transcript,
+        )
+    except MediaUnsupportedInputError as exc:
+        raise UnsupportedInputError(str(exc)) from exc
+    except MediaProcessingFailedError as exc:
+        raise ProcessingFailedError(str(exc)) from exc
+
+    return (
+        _text_context_from_markdown(
+            result.markdown,
+            fallback_title=result.title,
+            title=result.title,
+            max_chunk_size=40000,
+        ),
+        result.platform,
+    )

aimd/platform_utils.py

@@ -0,0 +1,26 @@
+"""Platform and hardware detection helpers."""
+
+from functools import lru_cache
+import platform
+import subprocess
+
+
+@lru_cache(maxsize=1)
+def is_apple_silicon() -> bool:
+    """Return True when running on Apple Silicon macOS."""
+    if platform.system() != "Darwin":
+        return False
+    try:
+        result = subprocess.run(
+            ["sysctl", "-n", "machdep.cpu.brand_string"],
+            capture_output=True,
+            text=True,
+            check=True,
+        )
+    except (subprocess.SubprocessError, FileNotFoundError):
+        return False
+
+    cpu_info = result.stdout.strip().lower()
+    return "apple" in cpu_info and any(
+        chip in cpu_info for chip in ("m1", "m2", "m3", "m4")
+    )

aimd/types.py

@@ -0,0 +1,12 @@
+from pydantic import BaseModel, Field
+
+
+class TextContext(BaseModel):
+    """Context for text processing with title and content."""
+
+    title: str = Field(..., description="Title of the text")
+    chunk_list: list[str] = Field(..., description="List of combined text chunks")
+    split_header_level: int | None = Field(
+        default=None,
+        description="Header level used for splitting (1-6), None if no splitting was done",
+    )

aimd/utils.py

@@ -0,0 +1,70 @@
+import re
+from pathlib import Path
+from urllib.parse import urlparse
+
+
+def sanitize_filename(title: str, max_length: int = 100) -> str:
+    """Sanitize title for use as filename.
+
+    Args:
+        title: Title to sanitize
+        max_length: Maximum length for filename
+
+    Returns:
+        Sanitized filename safe for filesystem
+    """
+
+    # Remove or replace invalid characters
+    sanitized = re.sub(r'[<>:"/\|?*]', "_", title)
+    # Remove extra whitespace and replace with underscores
+    sanitized = re.sub(r"\s+", "_", sanitized.strip())
+    # Remove leading/trailing dots and underscores
+    sanitized = sanitized.strip("._")
+    # Limit length
+    if len(sanitized) > max_length:
+        sanitized = sanitized[:max_length].rstrip("._")
+    # Ensure we have a valid filename
+    if not sanitized:
+        sanitized = "output"
+    return sanitized
+
+
+def create_output_path_from_title(
+    title: str, template_name: str, current_dir: Path = None
+) -> Path:
+    """Create output path using title and template name.
+
+    Args:
+        title: Title from TextContext
+        template_name: Template name for suffix
+        current_dir: Directory to save file (defaults to current working directory)
+
+    Returns:
+        Output path with sanitized title and template suffix
+    """
+    if current_dir is None:
+        current_dir = Path.cwd()
+
+    sanitized_title = sanitize_filename(title)
+    filename = f"{sanitized_title}_{template_name}.md"
+    return current_dir / filename
+
+
+def is_url(s: str) -> bool:
+    """Check if a string is a URL using basic URL parsing.
+
+    Args:
+        s: String to check
+
+    Returns:
+        True if string appears to be a URL
+    """
+    try:
+        result = urlparse(s)
+        # A non-empty scheme and netloc are strong indicators of a URL.
+        # We check for scheme presence, and also for netloc to catch schemeless URLs like "www.google.com".
+        return all([result.scheme, result.netloc]) or (
+            result.scheme in ["http", "https"] and not result.netloc
+        )
+    except ValueError:
+        return False

aimd_cli-0.9.2.dist-info/METADATA

@@ -0,0 +1,23 @@
+Metadata-Version: 2.3
+Name: aimd-cli
+Version: 0.9.2
+Summary: Context preparation tool for LLM workflows.
+Author: Shu Li
+Author-email: Shu Li <zetarylee@gmail.com>
+Requires-Dist: aimd-media
+Requires-Dist: aimd-book
+Requires-Dist: logly>=0.1.6
+Requires-Dist: markitdown>=0.1.1,<0.2.0
+Requires-Dist: pydantic>=2.0.0
+Requires-Dist: python-dotenv>=1.1.1
+Requires-Dist: rich>=13.9.4
+Requires-Dist: typer>=0.20.0
+Requires-Dist: aimd-api ; extra == 'all'
+Requires-Dist: aimd-mcp ; extra == 'all'
+Requires-Dist: aimd-ocr ; extra == 'all'
+Requires-Dist: aimd-html ; extra == 'all'
+Requires-Python: >=3.10, <3.13
+Provides-Extra: all
+Description-Content-Type: text/markdown
+
+Prepare LLM-ready context from URLs, audio/video, and documents.

aimd_cli-0.9.2.dist-info/RECORD

@@ -0,0 +1,39 @@
+aimd/AGENTS.md,sha256=qWZdW57zt9dmCADPdDbSjmxfyH7pWMeECtA3PUpuGq0,3954
+aimd/__init__.py,sha256=mKPHHwULavRIzkNJ9htoTO9FO_8TgS__g6E0_BpaD0A,241
+aimd/adapters/AGENTS.md,sha256=l1MEcwqziBKC28oXJsamh2vDTh3sdpKQbHSfbUXy84E,982
+aimd/adapters/__init__.py,sha256=kPCiAvKFPgizgLXLTrC4RW1oZkbMRW5SFLoTnBCJHZA,32
+aimd/adapters/cli/__init__.py,sha256=R6ccDx1BPyu31QO3lmw57_eCke4CkeAZAuEbGLDk96s,27
+aimd/adapters/cli/app.py,sha256=-MS60tmFkGDJtd50CcxE16i2vAVq0br96xt5nlf2gsM,7465
+aimd/application/AGENTS.md,sha256=LAeM0Jm2-HsAXJbNqqNO-wfX-CZ8elGrIXJn7KioUy4,2046
+aimd/application/__init__.py,sha256=fmMLANSxZPj3RKJ4srXjHEDG0ZgDvcTjLeuv1TDpKi0,320
+aimd/application/bootstrap.py,sha256=JJGZK-Il4vo5-cpQqCqZxwi9sa3Xn3GZujdHGTUKI6Q,1650
+aimd/application/models.py,sha256=PGmEkiKaGLYz7EsOneWZioSAfaVR1A3NxNkJX4qcQKA,1122
+aimd/application/services/__init__.py,sha256=7O6qLl42-prqJ_SujBpJffj3T9__V6pLXHu49TEw-E4,35
+aimd/application/services/interface_payloads.py,sha256=IwJViNVoUu06FkVhETQ9lt59RFcIh1lm9LtGYO1Q1vE,2528
+aimd/application/services/output_writer.py,sha256=PCoYRK-B-Ul6JoI1SAA7_YDYVgTc0_XMWbWytALrmUQ,2033
+aimd/application/use_cases/__init__.py,sha256=Tv3oOFjorCm6Etffgl4HsIUrmgtqgc6-c_OLQxGa8Ow,29
+aimd/application/use_cases/input_routing.py,sha256=7gvv03tSmvEz1oNr2AIU3BIOQPxsGwUXxI7dZkDw1G0,1946
+aimd/application/use_cases/list_engines.py,sha256=QIVjRguTA2NKEWlitaBn9AN-BzW6g6_aweSqiTc7rKk,1000
+aimd/application/use_cases/process_input.py,sha256=Q7UzOxmAj8Cl048o_uDpB0GECeEtm6cyzlc9c_FCxHY,1540
+aimd/application/use_cases/processors/__init__.py,sha256=XXYSDDR3lYRxmyRSEw-lsB4xEejc7mVPg_cK0CPFgow,365
+aimd/application/use_cases/processors/_base.py,sha256=1B5BjSzhrP6CNhDqVOcGfwj4dGooQHmOvqmRDAEifzA,415
+aimd/application/use_cases/processors/convert.py,sha256=m5VBEMjmlVUjJDpwQGs477HdHRvne0uSFuNmqHMPRQ4,974
+aimd/application/use_cases/processors/transcript.py,sha256=SMyX4H2nJZNM8V1Wotw-uQdgIgoVy1kIwPQGShgDV5U,2630
+aimd/cli.py,sha256=iJYbgVudhscnwgFV51DgtVXP-uC_OtEnMqwSmMdUBHc,130
+aimd/const.py,sha256=8ptoja8E7lgOdorDqgG3mF12oI81fykHrfhLaiOYxcg,724
+aimd/errors.py,sha256=pjMg0qPUZmly7nvke0MDYoZGp0-Gm2PD3rfbdjHScsw,902
+aimd/infrastructure/AGENTS.md,sha256=OHIS53QxMfo9bABNqsFo6PYfIuqKYrZNPYS5J6VYFaU,1782
+aimd/infrastructure/__init__.py,sha256=D_StCHKRqyvxeKpP-_lrddfbBbJe75K86cT5biZvmcw,36
+aimd/infrastructure/documents/__init__.py,sha256=efTOlhTS6Yl054gdPxPIKlyyRDTgbHiVrPvZkb4-Svk,534
+aimd/infrastructure/documents/chunking.py,sha256=3e-JylggA0qxYlCFGsy2DNC9uSSSrUV_bwiTWtcmGRU,5542
+aimd/infrastructure/documents/title_extractor.py,sha256=9dCjBoP-0PRRiXutTxyEnzMnxGoCbHy6rFZ5KL3W4UE,3215
+aimd/infrastructure/markitdown_processor.py,sha256=K8jajNE27KdMToLLI3LRol_3zh6DuO4sdgeIMuiDRHE,3224
+aimd/infrastructure/media_processor.py,sha256=lDCBJvx0cVL0N17f8OK9CjmXUirdsdnJA4LhqaM-gZ8,1745
+aimd/platform_utils.py,sha256=uM-Mj2fXLpZVoh-r0Ywe8xYNKgo-oEyrD7DqIeBSWDg,713
+aimd/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+aimd/types.py,sha256=OpedoG-3X5v-xIwKxKH6zxe0cebj6gzI2DafVVQFLNU,441
+aimd/utils.py,sha256=lc0uQQVK3RSyN2fAXLyMU-Psbi3gWQT0lhj8fRFc7vA,2076
+aimd_cli-0.9.2.dist-info/WHEEL,sha256=eh7sammvW2TypMMMGKgsM83HyA_3qQ5Lgg3ynoecH3M,79
+aimd_cli-0.9.2.dist-info/entry_points.txt,sha256=Q5kTy2HVVf4rZGtsnhXxsabRQ8uX1ZloE0hIacd4fH0,40
+aimd_cli-0.9.2.dist-info/METADATA,sha256=iuO6q7aLUaWhYqYNKTjjIdMU_u24A7Sd6tV5IbwOcIQ,727
+aimd_cli-0.9.2.dist-info/RECORD,,

aimd_cli-0.9.2.dist-info/WHEEL

@@ -0,0 +1,4 @@
+Wheel-Version: 1.0
+Generator: uv 0.8.24
+Root-Is-Purelib: true
+Tag: py3-none-any

aimd_cli-0.9.2.dist-info/entry_points.txt

@@ -0,0 +1,3 @@
+[console_scripts]
+aimd = aimd.cli:main
+