"""Extract one selectable-text PDF locally. See README before using scanned PDFs."""
from pathlib import Path
import argparse
import importlib.metadata
import json
import time
from docling.datamodel.base_models import InputFormat
from docling.datamodel.pipeline_options import PdfPipelineOptions
from docling.datamodel.accelerator_options import AcceleratorOptions, AcceleratorDevice
from docling.document_converter import DocumentConverter, PdfFormatOption

parser=argparse.ArgumentParser(description=__doc__)
parser.add_argument('--input',type=Path,default=Path('supplier-catalog.pdf'))
parser.add_argument('--output',type=Path,default=Path('extracted'))
args=parser.parse_args()
if not args.input.is_file():
    parser.error(f'Input PDF not found: {args.input}')
args.output.mkdir(parents=True,exist_ok=True)

options=PdfPipelineOptions()
options.do_ocr=False
options.do_table_structure=True
options.enable_remote_services=False
options.accelerator_options=AcceleratorOptions(device=AcceleratorDevice.CPU,num_threads=4)
converter=DocumentConverter(format_options={InputFormat.PDF:PdfFormatOption(pipeline_options=options)})
started=time.time()
result=converter.convert(args.input)
doc=result.document
(args.output/'supplier-catalog.md').write_text(doc.export_to_markdown(),encoding='utf-8')
tables=[]
for index,table in enumerate(doc.tables,start=1):
    frame=table.export_to_dataframe(doc=doc)
    filename=f'supplier-table-{index}.csv'
    frame.to_csv(args.output/filename,index=False)
    tables.append({'filename':filename,'rows':len(frame),'columns':list(map(str,frame.columns))})
receipt={'doclingVersion':importlib.metadata.version('docling'),'source':str(args.input),'conversionStatus':str(result.status),'pages':len(doc.pages),'tables':tables,'ocrEnabled':False,'remoteServicesEnabled':False,'conversionSecondsIncludingModelInitialization':round(time.time()-started,2),'exampleScope':'Small digitally created PDF; not an OCR or general accuracy benchmark.'}
(args.output/'conversion-receipt.json').write_text(json.dumps(receipt,indent=2)+'\n')
print(json.dumps(receipt,indent=2),flush=True)