src/file_parser.py

"""
File parser for Offer Catcher.

Supports .txt, .md, .pdf, and .docx.
Pure-local parsing: no file is saved to disk and no API call is made.
"""

from typing import Tuple


def extract_text_from_upload(uploaded_file) -> Tuple[str, str]:
    """
    Read a Streamlit UploadedFile and extract text.

    Returns:
        (text, message)
        - text: extracted resume text, or "" on failure
        - message: UI-friendly status message
    """
    if uploaded_file is None:
        return "", "未上传文件。"

    filename = getattr(uploaded_file, "name", "unknown")
    suffix = ""
    if "." in filename:
        suffix = "." + filename.rsplit(".", 1)[-1].lower()

    if suffix in (".txt", ".md"):
        return _read_text_file(uploaded_file, filename)
    if suffix == ".pdf":
        return _read_pdf(uploaded_file, filename)
    if suffix == ".docx":
        return _read_docx(uploaded_file, filename)
    if suffix == ".doc":
        return (
            "",
            f"不支持 .doc 格式（{filename}）。\n"
            "请先将文件另存为 .docx 或 PDF，再重新上传。",
        )

    return "", f"不支持的文件格式：{suffix}（{filename}）。"


def _reset_file(uploaded_file) -> None:
    if hasattr(uploaded_file, "seek"):
        try:
            uploaded_file.seek(0)
        except Exception:
            pass


def _read_text_file(uploaded_file, filename: str) -> Tuple[str, str]:
    raw = uploaded_file.read()
    _reset_file(uploaded_file)

    if not raw:
        return "", f"文件为空（{filename}）。"

    text = None
    for encoding in ("utf-8", "utf-8-sig", "gbk", "gb18030", "latin-1"):
        try:
            text = raw.decode(encoding)
            break
        except UnicodeDecodeError:
            continue

    if text is None:
        return "", f"无法识别文件编码（{filename}）。请另存为 UTF-8 文本后重试。"

    text = text.strip()
    if len(text) < 30:
        return "", f"文件内容过少（仅 {len(text)} 字符，{filename}），请检查文件是否正确。"

    return text, f"已解析 {filename}（{len(text)} 字符，文本格式）。"


def _read_pdf(uploaded_file, filename: str) -> Tuple[str, str]:
    try:
        import pypdf
    except ImportError:
        return (
            "",
            f"服务器未安装 pypdf，无法解析 PDF（{filename}）。\n"
            "请在 requirements.txt 中加入 pypdf>=4.0.0 并重新部署。",
        )

    try:
        _reset_file(uploaded_file)
        reader = pypdf.PdfReader(uploaded_file)
        parts = []
        for page in reader.pages:
            try:
                page_text = page.extract_text()
            except Exception:
                page_text = ""
            if page_text:
                parts.append(page_text)
        text = "\n".join(parts).strip()

        if len(text) < 30:
            return (
                "",
                f"PDF 解析结果过少（仅 {len(text)} 字符，{filename}）。\n"
                "可能是扫描版 PDF，请尝试复制文本后粘贴，或上传 DOCX/PDF 文字版。",
            )

        return text, f"已解析 {filename}（{len(text)} 字符，PDF {len(reader.pages)} 页）。"
    except Exception as exc:
        return "", f"PDF 解析失败（{filename}）：{exc}"


def _read_docx(uploaded_file, filename: str) -> Tuple[str, str]:
    try:
        import docx
    except ImportError:
        return (
            "",
            f"服务器未安装 python-docx，无法解析 DOCX（{filename}）。\n"
            "请在 requirements.txt 中加入 python-docx>=1.1.0 并重新部署。",
        )

    try:
        _reset_file(uploaded_file)
        document = docx.Document(uploaded_file)
        parts = []

        for paragraph in document.paragraphs:
            value = paragraph.text.strip()
            if value:
                parts.append(value)

        for table in document.tables:
            for row in table.rows:
                cells = [cell.text.strip() for cell in row.cells if cell.text.strip()]
                if cells:
                    parts.append(" | ".join(cells))

        text = "\n".join(parts).strip()
        if len(text) < 30:
            return "", f"DOCX 解析结果过少（仅 {len(text)} 字符，{filename}），请检查文件是否正确。"

        return text, f"已解析 {filename}（{len(text)} 字符，DOCX）。"
    except Exception as exc:
        return "", f"DOCX 解析失败（{filename}）：{exc}"