ordinaceprojekt/RozkladScanu/navrh_pojmenovani.py

"""
navrh_pojmenovani.py — Claude navrhne jmeno souboru pro naskenovanou brozuru.

Vezme prvnich N stranek PDF, posle je Claudovi jako obrazky a navrhne nazev.
Dialog umozni opravit navrh. Historie dvojic (navrh Claudea, volba uzivatele)
se ulozi a priste prikladem uci Claudea pojmenovavat podle tvych preferencí.

Pouziti:
    python navrh_pojmenovani.py vystup/brozura.pdf
    python navrh_pojmenovani.py vystup/brozura.pdf vystup/muj_nazev.pdf
"""

import sys
import json
import base64
import io
import copy
import tkinter as tk
from pathlib import Path
from datetime import date
import fitz          # pymupdf
import anthropic

HISTORY_FILE = Path(__file__).parent / "pojmenovani_historie.json"
N_PAGES      = 5
DPI          = 120   # dostatecne pro cteni obsahu, nizke naklady na tokeny


# ---------- pomocne funkce ----------

def load_history() -> list:
    if HISTORY_FILE.exists():
        return json.loads(HISTORY_FILE.read_text(encoding="utf-8"))
    return []


def save_history(history: list):
    HISTORY_FILE.write_text(
        json.dumps(history, ensure_ascii=False, indent=2),
        encoding="utf-8",
    )


def pdf_pages_to_b64(pdf_path: str, n: int = N_PAGES) -> list[str]:
    """Vrati seznam base64 JPEG retezcu pro prvnich n stranek PDF."""
    doc = fitz.open(pdf_path)
    pages = min(n, len(doc))
    result = []
    mat = fitz.Matrix(DPI / 72, DPI / 72)
    for i in range(pages):
        pix = doc[i].get_pixmap(matrix=mat, colorspace=fitz.csRGB)
        jpeg = pix.tobytes("jpeg", jpg_quality=75)
        result.append(base64.standard_b64encode(jpeg).decode())
    doc.close()
    return result


def ask_claude(images_b64: list[str], history: list) -> str:
    """Posle stranky Claudovi a vrati navrzeny nazev souboru."""
    client = anthropic.Anthropic()

    history_text = ""
    if history:
        history_text = "\n\nPriklady z minulosti (muj navrh → uzivatel zvolil):\n"
        for item in history[-15:]:
            history_text += f"  {item['claude']} -> {item['user']}\n"

    content: list = []
    for i, b64 in enumerate(images_b64):
        content.append({
            "type": "image",
            "source": {"type": "base64", "media_type": "image/jpeg", "data": b64},
        })

    today = date.today().strftime("%Y-%m-%d")
    content.append({
        "type": "text",
        "text": (
            f"Toto je prvnich {len(images_b64)} stranek naskenované brozury. "
            "Navrhni vhodny kratky nazev souboru (bez pripony). "
            f"Zacni vzdy datem dnesniho dne ve formatu YYYY-MM-DD ({today}), "
            "pak podtrzitko a popis obsahu. "
            "Pouzij pouze ASCII znaky, cisla, podtrzitka nebo pomlcky — zadna diakritika ani mezery. "
            "Popis ma vystihovat obsah dokumentu, byt strucny (2–4 slova)."
            f"{history_text}\n\n"
            "Odpovez pouze samotnym nazvem souboru, nic jineho. "
            f"Priklad spravneho formatu: {today}_nazev_dokumentu"
        ),
    })

    response = client.messages.create(
        model="claude-opus-4-7",
        max_tokens=64,
        thinking={"type": "adaptive"},
        messages=[{"role": "user", "content": content}],
    )

    for block in response.content:
        if block.type == "text":
            return block.text.strip().strip('"').strip("'")
    return "brozura"


def dialog(claude_suggestion: str) -> dict:
    """Zobrazi dialog s navrzenim nazvem, uzivatel muze opravit."""
    root = tk.Tk()
    root.title("Navrh pojmenovani")
    root.resizable(False, False)
    root.attributes("-topmost", True)

    result = {"name": None, "cancelled": False}

    tk.Label(
        root, text="Claude navrhuje:", font=("Segoe UI", 9, "italic"), fg="#555"
    ).pack(padx=24, pady=(16, 2), anchor="w")

    tk.Label(
        root, text=claude_suggestion, font=("Segoe UI", 12, "bold"), fg="#1a6aaa"
    ).pack(padx=24, anchor="w")

    tk.Label(
        root, text="Nazev souboru (uprav pokud chces):", font=("Segoe UI", 9)
    ).pack(padx=24, pady=(12, 2), anchor="w")

    entry = tk.Entry(root, width=52, font=("Segoe UI", 10))
    entry.insert(0, claude_suggestion)
    entry.pack(padx=24)
    entry.select_range(0, tk.END)
    entry.focus_set()

    def ok(event=None):
        val = entry.get().strip()
        if val:
            result["name"] = val
        root.destroy()

    def cancel(event=None):
        result["cancelled"] = True
        root.destroy()

    btn = tk.Frame(root)
    btn.pack(pady=16)
    tk.Button(btn, text="OK", width=12, command=ok).pack(side=tk.LEFT, padx=6)
    tk.Button(btn, text="Zrusit", width=12, command=cancel).pack(side=tk.LEFT, padx=6)

    root.bind("<Return>", ok)
    root.bind("<Escape>", cancel)
    root.mainloop()
    return result


# ---------- hlavni logika ----------

def main():
    if len(sys.argv) < 2:
        print(__doc__)
        sys.exit(1)

    pdf_path = Path(sys.argv[1])
    if not pdf_path.exists():
        print(f"Soubor nenalezen: {pdf_path}")
        sys.exit(1)

    # vystupni cesta — volitelny druhy argument
    out_path: Path | None = None
    if len(sys.argv) > 2:
        out_path = Path(sys.argv[2])

    print(f"Nacitam stranky z: {pdf_path}")
    images = pdf_pages_to_b64(str(pdf_path), N_PAGES)
    print(f"  {len(images)} stranek prevedeno na obrazky")

    history = load_history()
    print(f"  {len(history)} polozek v historii pojmenovani")

    print("Ptam se Claudea...")
    claude_name = ask_claude(images, history)
    print(f"  Claude navrhuje: {claude_name}")

    res = dialog(claude_name)

    if res["cancelled"] or not res["name"]:
        print("Zruseno — soubor nebyl prejmenovan.")
        return

    user_name = res["name"]

    # uloz do historie
    history.append({"claude": claude_name, "user": user_name})
    save_history(history)
    print(f"  Historie ulozena ({len(history)} polozek)")

    # urceni vystupniho souboru
    if out_path is None:
        out_path = pdf_path.parent / (user_name + pdf_path.suffix)

    if out_path != pdf_path:
        pdf_path.rename(out_path)
        print(f"Prejmenovan: {out_path}")
    else:
        print(f"Nazev beze zmeny: {out_path}")


if __name__ == "__main__":
    main()