customkun_any / glmocr.py
showeed's picture
Upload 8 files
11bcfa7 verified
"""
GLM-OCR ๆฑŽ็”จ็”ปๅƒOCRใƒปExcelๅ‡บๅŠ›ใ‚นใ‚ฏใƒชใƒ—ใƒˆ
ใƒขใƒ‡ใƒซ : zai-org/GLM-OCR (HuggingFace)
่จญๅฎš : YAML ใพใŸใฏ Excel (.xlsx) ใฎใ‚ณใƒณใƒ•ใ‚ฃใ‚ฐใƒ•ใ‚กใ‚คใƒซใงๆŠฝๅ‡บ้ …็›ฎใƒป็”ปๅƒใƒปๅ‡บๅŠ›ๅ…ˆใ‚’ๅฎš็พฉ
ไฝฟใ„ๆ–น :
python glmocr.py --config configs/invoice.yaml
python glmocr.py --config configs/invoice.xlsx
python glmocr.py --config configs/invoice.yaml --image scan.pdf
python glmocr.py --config configs/invoice.yaml --create-excel # Excel ใƒ†ใƒณใƒ—ใƒฌใƒผใƒˆ็”Ÿๆˆ
ๅ‡บๅŠ› : {output_dir}/{configๅ}.xlsx๏ผˆใ‚ปใ‚ฏใ‚ทใƒงใƒณใ”ใจใซใ‚ทใƒผใƒˆใ‚’ๅˆ†ใ‘ใฆไฟๅญ˜๏ผ‰
PDF ่ค‡ๆ•ฐใƒšใƒผใ‚ธใฎๅ ดๅˆใฏใ‚ทใƒผใƒˆๅใ‚’ P01_/P02_... ใงใƒšใƒผใ‚ธๅŒบๅˆฅใ™ใ‚‹
"""
import argparse
import json
import re
import sys
from html.parser import HTMLParser
from pathlib import Path
from config_loader import load_config, create_excel_template
from preprocess import apply_preprocess, load_input_images
# Windows ใ‚ณใƒณใ‚ฝใƒผใƒซใฎๆ–‡ๅญ—ๅŒ–ใ‘ๅฏพ็ญ–
if sys.stdout.encoding != "utf-8":
sys.stdout.reconfigure(encoding="utf-8", errors="replace")
if sys.stderr.encoding != "utf-8":
sys.stderr.reconfigure(encoding="utf-8", errors="replace")
import pandas as pd
import torch
from transformers import AutoProcessor, AutoModelForImageTextToText
MODEL_ID = "zai-org/GLM-OCR"
# โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€
# JSON ใ‚นใ‚ญใƒผใƒžๅ‹•็š„็”Ÿๆˆ
# โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€
def build_json_schema(sections: dict) -> str:
"""YAML sections ๅฎš็พฉใ‹ใ‚‰ GLM-OCR ็”จ JSON ใ‚นใ‚ญใƒผใƒžๆ–‡ๅญ—ๅˆ—ใ‚’็”Ÿๆˆใ™ใ‚‹ใ€‚
Args:
sections: YAML ใฎ sections ่พžๆ›ธ
Returns:
str: JSON ใ‚นใ‚ญใƒผใƒžๆ–‡ๅญ—ๅˆ—
"""
schema = {name: cfg["fields"] for name, cfg in sections.items()}
return json.dumps(schema, ensure_ascii=False, indent=2)
# โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€
# ใƒขใƒ‡ใƒซ่ชญใฟ่พผใฟ
# โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€
def load_model():
"""GLM-OCR ใƒขใƒ‡ใƒซใจ Processor ใ‚’่ชญใฟ่พผใ‚€ใ€‚
Returns:
tuple[AutoModelForImageTextToText, AutoProcessor]: ใƒขใƒ‡ใƒซใจใƒ—ใƒญใ‚ปใƒƒใ‚ต
"""
print(f"[INFO] ใƒขใƒ‡ใƒซใ‚’่ชญใฟ่พผใ‚“ใงใ„ใพใ™: {MODEL_ID}")
processor = AutoProcessor.from_pretrained(MODEL_ID)
model = AutoModelForImageTextToText.from_pretrained(
MODEL_ID,
torch_dtype="auto",
device_map="auto",
)
model.eval()
print(f"[INFO] ใƒขใƒ‡ใƒซ่ชญใฟ่พผใฟๅฎŒไบ† (device: {model.device})")
return model, processor
# โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€
# ๆŽจ่ซ–
# โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€
def run_ocr(model, processor, pil_image: Image.Image, prompt: str) -> str:
"""ๅ˜ไธ€ใƒ—ใƒญใƒณใƒ—ใƒˆใง GLM-OCR ๆŽจ่ซ–ใ‚’ๅฎŸ่กŒใ™ใ‚‹ใ€‚
Args:
model: GLM-OCR ใƒขใƒ‡ใƒซ
processor: GLM-OCR ใƒ—ใƒญใ‚ปใƒƒใ‚ต
pil_image: ๅ…ฅๅŠ›็”ปๅƒ (PIL.Image)
prompt: OCR ใƒ—ใƒญใƒณใƒ—ใƒˆๆ–‡ๅญ—ๅˆ—
Returns:
str: ใƒขใƒ‡ใƒซใŒ็”Ÿๆˆใ—ใŸใƒ†ใ‚ญใ‚นใƒˆ
"""
messages = [
{
"role": "user",
"content": [
{"type": "image", "image": pil_image},
{"type": "text", "text": prompt},
],
}
]
inputs = processor.apply_chat_template(
messages,
tokenize=True,
add_generation_prompt=True,
return_dict=True,
return_tensors="pt",
)
inputs.pop("token_type_ids", None)
inputs = {k: v.to(model.device) for k, v in inputs.items()}
with torch.no_grad():
generated_ids = model.generate(**inputs, max_new_tokens=8192)
output_text = processor.decode(
generated_ids[0][inputs["input_ids"].shape[1]:],
skip_special_tokens=True,
)
return output_text.strip()
# โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€
# ใƒ‘ใƒผใ‚น: HTML ใƒ†ใƒผใƒ–ใƒซ โ†’ DataFrame
# โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€
class _HtmlTableParser(HTMLParser):
"""HTML ใƒ†ใƒผใƒ–ใƒซใ‚’ใƒ‘ใƒผใ‚นใ—ใฆ่กŒใƒชใ‚นใƒˆใ‚’ๅŽ้›†ใ™ใ‚‹ใ‚ทใƒณใƒ—ใƒซใชใƒ‘ใƒผใ‚ตใƒผใ€‚"""
def __init__(self):
super().__init__()
self.rows: list[list[str]] = []
self._current_row: list[str] = []
self._current_cell: str = ""
self._in_cell: bool = False
def handle_starttag(self, tag, attrs):
if tag == "tr":
self._current_row = []
elif tag in ("td", "th"):
self._current_cell = ""
self._in_cell = True
def handle_endtag(self, tag):
if tag in ("td", "th"):
self._current_row.append(self._current_cell.strip())
self._in_cell = False
elif tag == "tr":
if self._current_row:
self.rows.append(self._current_row)
def handle_data(self, data):
if self._in_cell:
self._current_cell += data
def parse_html_table(text: str) -> pd.DataFrame:
"""OCR ๅ‡บๅŠ›ใƒ†ใ‚ญใ‚นใƒˆใ‹ใ‚‰ HTML ใƒ†ใƒผใƒ–ใƒซใ‚’ๆŠฝๅ‡บใ—ใฆ DataFrame ใซๅค‰ๆ›ใ™ใ‚‹ใ€‚
Args:
text: OCR ใƒขใƒ‡ใƒซใฎๅ‡บๅŠ›ใƒ†ใ‚ญใ‚นใƒˆ
Returns:
pd.DataFrame: ใƒ†ใƒผใƒ–ใƒซใƒ‡ใƒผใ‚ฟใ€‚่ฆ‹ใคใ‹ใ‚‰ใชใ„ๅ ดๅˆใฏ็ฉบใฎ DataFrameใ€‚
"""
match = re.search(r"<table.*?>.*?</table>", text, re.DOTALL | re.IGNORECASE)
if not match:
return pd.DataFrame()
parser = _HtmlTableParser()
parser.feed(match.group(0))
if len(parser.rows) < 2:
return pd.DataFrame()
return pd.DataFrame(parser.rows[1:], columns=parser.rows[0])
# โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€
# ใƒ‘ใƒผใ‚น: Markdown ใƒ†ใƒผใƒ–ใƒซ โ†’ DataFrame
# โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€
def parse_markdown_table(text: str) -> pd.DataFrame:
"""OCR ๅ‡บๅŠ›ใƒ†ใ‚ญใ‚นใƒˆใ‹ใ‚‰ Markdown ใƒ†ใƒผใƒ–ใƒซใ‚’ๆŠฝๅ‡บใ—ใฆ DataFrame ใซๅค‰ๆ›ใ™ใ‚‹ใ€‚
Args:
text: OCR ใƒขใƒ‡ใƒซใฎๅ‡บๅŠ›ใƒ†ใ‚ญใ‚นใƒˆ
Returns:
pd.DataFrame: ใƒ†ใƒผใƒ–ใƒซใƒ‡ใƒผใ‚ฟใ€‚่ฆ‹ใคใ‹ใ‚‰ใชใ„ๅ ดๅˆใฏ็ฉบใฎ DataFrameใ€‚
"""
table_lines = [l for l in text.splitlines() if "|" in l]
if len(table_lines) < 2:
return pd.DataFrame()
data_lines = [l for l in table_lines if not re.match(r"^\|[\s\-:|]+\|$", l)]
rows = [[c.strip() for c in l.strip().strip("|").split("|")] for l in data_lines]
if not rows:
return pd.DataFrame()
return pd.DataFrame(rows[1:], columns=rows[0])
def parse_table(text: str) -> pd.DataFrame:
"""HTML ใพใŸใฏ Markdown ใƒ†ใƒผใƒ–ใƒซใ‚’่‡ชๅ‹•ๅˆคๅˆฅใ—ใฆใƒ‘ใƒผใ‚นใ™ใ‚‹ใ€‚
Args:
text: OCR ใƒขใƒ‡ใƒซใฎๅ‡บๅŠ›ใƒ†ใ‚ญใ‚นใƒˆ
Returns:
pd.DataFrame: ใƒ†ใƒผใƒ–ใƒซใƒ‡ใƒผใ‚ฟใ€‚่ฆ‹ใคใ‹ใ‚‰ใชใ„ๅ ดๅˆใฏ็ฉบใฎ DataFrameใ€‚
"""
if "<table" in text.lower():
return parse_html_table(text)
return parse_markdown_table(text)
# โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€
# ใƒ‘ใƒผใ‚น: JSON ใƒ†ใ‚ญใ‚นใƒˆ โ†’ dict
# โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€
def parse_json_output(text: str) -> dict:
"""OCR ๅ‡บๅŠ›ใƒ†ใ‚ญใ‚นใƒˆใ‹ใ‚‰ JSON ้ƒจๅˆ†ใ‚’ๆŠฝๅ‡บใ—ใฆใƒ‘ใƒผใ‚นใ™ใ‚‹ใ€‚
Args:
text: OCR ใƒขใƒ‡ใƒซใฎๅ‡บๅŠ›ใƒ†ใ‚ญใ‚นใƒˆ
Returns:
dict: ใƒ‘ใƒผใ‚นใ•ใ‚ŒใŸ JSON ใƒ‡ใƒผใ‚ฟใ€‚ๅคฑๆ•—ๆ™‚ใฏ็ฉบใฎ dictใ€‚
"""
code_block = re.search(r"```(?:json)?\s*(\{.*?\})\s*```", text, re.DOTALL)
json_str = code_block.group(1) if code_block else None
if not json_str:
brace_match = re.search(r"\{.*\}", text, re.DOTALL)
if not brace_match:
return {}
json_str = brace_match.group(0)
try:
return json.loads(json_str)
except json.JSONDecodeError:
json_str_fixed = re.sub(r",\s*([}\]])", r"\1", json_str)
try:
return json.loads(json_str_fixed)
except json.JSONDecodeError:
return {}
# โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€
# Excel ไฟๅญ˜๏ผˆๅ…จใ‚ทใƒผใƒˆใพใจใ‚ๆ›ธใ๏ผ‰
# โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€
def save_excel(sheets: dict[str, pd.DataFrame], filepath: Path) -> None:
"""่ค‡ๆ•ฐใฎ DataFrame ใ‚’ 1 ใคใฎ Excel ใƒ•ใ‚กใ‚คใƒซใซใ‚ทใƒผใƒˆใ”ใจใซไฟๅญ˜ใ™ใ‚‹ใ€‚
Args:
sheets: {ใ‚ทใƒผใƒˆๅ: DataFrame} ใฎ่พžๆ›ธ๏ผˆ็ฉบใฎ DataFrame ใฏ็ฉบใ‚ทใƒผใƒˆใจใ—ใฆไฟๅญ˜๏ผ‰
filepath: ๅ‡บๅŠ›ๅ…ˆ Excel ใƒ•ใ‚กใ‚คใƒซใฎใƒ‘ใ‚น (.xlsx)
"""
filepath.parent.mkdir(parents=True, exist_ok=True)
with pd.ExcelWriter(filepath, engine="openpyxl") as writer:
for sheet_name, df in sheets.items():
# Excel ใ‚ทใƒผใƒˆๅใฏ 31 ๆ–‡ๅญ—ไปฅๅ†…ใฎๅˆถ้™ใ‚ใ‚Š
safe_name = sheet_name[:31]
df.to_excel(writer, sheet_name=safe_name, index=False)
row_info = f"{len(df)} ่กŒ" if not df.empty else "ใƒ‡ใƒผใ‚ฟใชใ—"
print(f"[OK] ใ‚ทใƒผใƒˆ '{safe_name}' ใ‚’ๆ›ธใ่พผใฟใพใ—ใŸ ({row_info})")
print(f"[OK] Excel ไฟๅญ˜ๅฎŒไบ†: {filepath}")
# โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€
# ใ‚ปใ‚ฏใ‚ทใƒงใƒณ dict โ†’ DataFrame ๅค‰ๆ›
# โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€
def section_to_df(section: dict) -> pd.DataFrame:
"""1 ใƒฌใƒ™ใƒซใฎ dict ใ‚’ใ€Œkey / valueใ€ใฎ 2 ๅˆ— DataFrame ใซๅค‰ๆ›ใ™ใ‚‹ใ€‚
Args:
section: ใ‚ญใƒผใจๅ€คใ‚’ๆŒใค่พžๆ›ธ
Returns:
pd.DataFrame: key / value ใฎ 2 ๅˆ— DataFrame
"""
if not section:
return pd.DataFrame()
return pd.DataFrame({"key": list(section.keys()), "value": list(section.values())})
# โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€
# ใƒกใ‚คใƒณ
# โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€
def main():
"""ใƒกใ‚คใƒณๅ‡ฆ็†: ใ‚ณใƒณใƒ•ใ‚ฃใ‚ฐ่ชญใฟ่พผใฟ โ†’ ็”ปๅƒOCR โ†’ Excel ๅ‡บๅŠ›ใ€‚"""
parser = argparse.ArgumentParser(description="GLM-OCR ๆฑŽ็”จ็”ปๅƒOCRใƒปCSVๅ‡บๅŠ›ใ‚นใ‚ฏใƒชใƒ—ใƒˆ")
parser.add_argument(
"--config", "-c", required=True, type=Path,
help="ใ‚ณใƒณใƒ•ใ‚ฃใ‚ฐใƒ•ใ‚กใ‚คใƒซใฎใƒ‘ใ‚น๏ผˆ.yaml ใพใŸใฏ .xlsx๏ผ‰ไพ‹: configs/invoice.yaml",
)
parser.add_argument(
"--image", "-i", type=Path, default=None,
help="็”ปๅƒใƒ•ใ‚กใ‚คใƒซใฎใƒ‘ใ‚น๏ผˆ็œ็•ฅๆ™‚ใฏใ‚ณใƒณใƒ•ใ‚ฃใ‚ฐใฎ image ่จญๅฎšใ‚’ไฝฟ็”จ๏ผ‰",
)
parser.add_argument(
"--create-excel", action="store_true",
help="ใ‚ณใƒณใƒ•ใ‚ฃใ‚ฐใ‚’่ชญใฟ่พผใ‚“ใง Excel ใƒ†ใƒณใƒ—ใƒฌใƒผใƒˆใ‚’็”Ÿๆˆใ—ใฆ็ต‚ไบ†ใ™ใ‚‹",
)
args = parser.parse_args()
# โ”€โ”€ ใ‚ณใƒณใƒ•ใ‚ฃใ‚ฐ่ชญใฟ่พผใฟ โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€
config_path = args.config.resolve()
if not config_path.exists():
print(f"[ERROR] ใ‚ณใƒณใƒ•ใ‚ฃใ‚ฐใŒ่ฆ‹ใคใ‹ใ‚Šใพใ›ใ‚“: {config_path}", file=sys.stderr)
sys.exit(1)
cfg = load_config(config_path)
# โ”€โ”€ Excel ใƒ†ใƒณใƒ—ใƒฌใƒผใƒˆ็”Ÿๆˆใƒขใƒผใƒ‰ โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€
if args.create_excel:
xlsx_path = config_path.with_suffix(".xlsx")
create_excel_template(cfg, xlsx_path)
print(f"[INFO] Excel ใƒ†ใƒณใƒ—ใƒฌใƒผใƒˆใ‚’็”Ÿๆˆใ—ใพใ—ใŸ: {xlsx_path}")
sys.exit(0)
config_dir = config_path.parent.parent # configs/ ใฎ่ฆช = ใ‚นใ‚ฏใƒชใƒ—ใƒˆใฎใƒ‡ใ‚ฃใƒฌใ‚ฏใƒˆใƒช
# ็”ปๅƒใƒ‘ใ‚นใฎ่งฃๆฑบ๏ผˆCLIๅผ•ๆ•ฐ > YAML่จญๅฎš๏ผ‰
if args.image:
image_path = args.image.resolve()
else:
image_path = (config_dir / cfg["image"]).resolve()
output_dir = (config_dir / cfg.get("output_dir", "output")).resolve()
extract_table: bool = cfg.get("extract_table", True)
sections: dict = cfg.get("sections", {})
if not image_path.exists():
print(f"[ERROR] ็”ปๅƒใŒ่ฆ‹ใคใ‹ใ‚Šใพใ›ใ‚“: {image_path}", file=sys.stderr)
sys.exit(1)
# ๅ‡บๅŠ› Excel ใƒ•ใ‚กใ‚คใƒซๅ: {configๅ}.xlsx
excel_path = output_dir / f"{config_path.stem}.xlsx"
preprocess_cfg: dict = cfg.get("preprocess", {})
print(f"[INFO] ใ‚ณใƒณใƒ•ใ‚ฃใ‚ฐ : {config_path.name}")
print(f"[INFO] ๅฏพ่ฑกใƒ•ใ‚กใ‚คใƒซ: {image_path}")
print(f"[INFO] ๅ‡บๅŠ›ๅ…ˆ : {excel_path}")
print(f"[INFO] ใƒ†ใƒผใƒ–ใƒซ่ช่ญ˜: {'ใ‚ใ‚Š' if extract_table else 'ใชใ—'}")
print(f"[INFO] ๆŠฝๅ‡บใ‚ปใ‚ฏใ‚ทใƒงใƒณ: {list(sections.keys())}")
print(f"[INFO] ๅ‰ๅ‡ฆ็†่จญๅฎš: {preprocess_cfg or 'ๅ…จใ‚นใƒ†ใƒƒใƒ— ON๏ผˆใƒ‡ใƒ•ใ‚ฉใƒซใƒˆ๏ผ‰'}")
# โ”€โ”€ ๅ…ฅๅŠ›่ชญใฟ่พผใฟ๏ผˆ็”ปๅƒ or PDF ๅ…จใƒšใƒผใ‚ธ๏ผ‰โ”€โ”€
print(f"\n[INFO] ใƒ•ใ‚กใ‚คใƒซใ‚’่ชญใฟ่พผใ‚“ใงใ„ใพใ™...")
raw_pages = load_input_images(image_path)
total_pages = len(raw_pages)
print(f"[INFO] ใƒšใƒผใ‚ธๆ•ฐ: {total_pages}")
# โ”€โ”€ ใƒขใƒ‡ใƒซ่ชญใฟ่พผใฟ โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€
model, processor = load_model()
# ๆ›ธใ่พผใ‚€ใ‚ทใƒผใƒˆใ‚’ๅŽ้›†ใ™ใ‚‹่พžๆ›ธ {ใ‚ทใƒผใƒˆๅ: DataFrame}
sheets: dict[str, pd.DataFrame] = {}
# โ”€โ”€ ๅ„ใƒšใƒผใ‚ธใ‚’ๅ‡ฆ็† โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€
for page_no, raw_image in enumerate(raw_pages, start=1):
# ่ค‡ๆ•ฐใƒšใƒผใ‚ธใฎๅ ดๅˆใฏใ‚ทใƒผใƒˆๅใซ P01_ / P02_ ... ใ‚’ไป˜ไธŽ
prefix = f"P{page_no:02d}_" if total_pages > 1 else ""
print(f"\n{'โ”€' * 50}")
print(f"[INFO] ใƒšใƒผใ‚ธ {page_no}/{total_pages} ใ‚’ๅ‡ฆ็†ไธญ...")
# ๅ‰ๅ‡ฆ็†
pil_image = apply_preprocess(raw_image, preprocess_cfg)
print(f"[INFO] ็”ปๅƒใ‚ตใ‚คใ‚บ: {pil_image.size}")
# โ”€โ”€ ๆŽจ่ซ–โ‘ : ใƒ†ใƒผใƒ–ใƒซ่ช่ญ˜๏ผˆใ‚ชใƒ—ใ‚ทใƒงใƒณ๏ผ‰โ”€โ”€
if extract_table:
print("[INFO] ๆŽจ่ซ–โ‘  ใƒ†ใƒผใƒ–ใƒซ่ช่ญ˜ ใ‚’ๅฎŸ่กŒไธญ...")
table_text = run_ocr(model, processor, pil_image, "Table Recognition:")
print("[RAW] ใƒ†ใƒผใƒ–ใƒซ่ช่ญ˜ ๅ‡บๅŠ›:")
print(table_text)
print()
sheets[f"{prefix}table"] = parse_table(table_text)
# โ”€โ”€ ๆŽจ่ซ–โ‘ก: ๆง‹้€ ๅŒ– JSON ๆŠฝๅ‡บ โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€
if sections:
print("[INFO] ๆŽจ่ซ–โ‘ก ๆง‹้€ ๅŒ– JSON ๆŠฝๅ‡บ ใ‚’ๅฎŸ่กŒไธญ...")
json_schema = build_json_schema(sections)
extract_prompt = (
"Extract all the following information from this image "
"and fill in the JSON template below. "
"Return only valid JSON, no extra text.\n\n"
+ json_schema
)
json_text = run_ocr(model, processor, pil_image, extract_prompt)
print("[RAW] JSON ๆŠฝๅ‡บ ๅ‡บๅŠ›:")
print(json_text)
print()
data = parse_json_output(json_text)
for section_name, section_cfg in sections.items():
label = f"{prefix}{section_cfg.get('label', section_name)}"
sheets[label] = section_to_df(data.get(section_name, {}))
# โ”€โ”€ Excel ไฟๅญ˜ โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€
print()
save_excel(sheets, excel_path)
# โ”€โ”€ ็ตๆžœใ‚ตใƒžใƒชใƒผ่กจ็คบ โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€โ”€
print("\n" + "=" * 60)
print(" ๅ‡บๅŠ›็ตๆžœใ‚ตใƒžใƒชใƒผ")
print("=" * 60)
for sheet_name, df in sheets.items():
print(f"\nโ–ผ {sheet_name}")
print(df.to_string(index=False) if not df.empty else " (ใƒ‡ใƒผใ‚ฟใชใ—)")
print("\n[INFO] ๅ…จๅ‡ฆ็†ใŒๅฎŒไบ†ใ—ใพใ—ใŸใ€‚")
if __name__ == "__main__":
main()