import debug
import cluster
import config
import pdfplumber
from multiprocessing.pool import ThreadPool
import pypdf
import queue
import subprocess

def main():
   de_column_ify()

def log(*args):
   print(*args, flush=True)

def de_column_ify():
   q = queue.Queue(maxsize=4 if config.PARALLEL else 0)
   with pdfplumber.open(config.INPUT) as pdf:
      cropped_pages = []
      with ThreadPool(4) as pool:
         for i in range(len(pdf.pages)):
            if config.PARALLEL:
               pool.apply_async(de_columnify_page, (q, config.INPUT, pdf.pages[i], ))
            else:
               de_columnify_page(q, config.INPUT, pdf.pages[i])
         for i in range(len(pdf.pages)):
            log("getting", i, "of", len(pdf.pages))
            got = q.get()
            for got_i in got[1]:
               cropped_pages.append((got[0], got_i.path))
               #if got_i.chars:
               #   debug.debug_show(debug.debug_im(got_i.page))

   log("merging", len(cropped_pages), "de-column-ified pages")
   cropped_pages = sorted(cropped_pages)
   writer = pypdf.PdfWriter()
   for cropped_page in cropped_pages:
      with open(cropped_page[1], "rb") as f:
         reader = pypdf.PdfReader(f)
         writer.add_page(reader.pages[0])
   log("dumping de-column-ified pages")
   output = f'{config.INPUT}.de-column-ified.pdf'
   with open(output, "wb") as f:
      writer.write(f)
   log(output)

def de_columnify_page(q, path, page):
   for _ in range(3):
      try:
         result = cluster.Chars(path, page.chars, page).divide_into_columns()
         log("putting", page.page_number, len(result))
         q.put((page.page_number, result))
         return
      except Exception as e:
         log(page.page_number, "encountered", e)
   raise Exception(f"failure for {page.page_number}")

def textify(page):
   lines = page.extract_text(layout=True).split("\n")
   leading_spaces = []
   for line in lines:
      leading_spaces.append(len(line) - len(line.lstrip()))
   median_leading_spaces = sorted(leading_spaces)[len(leading_spaces)//2]

   paragraphs = []
   for line in lines:
      if line.startswith(median_leading_spaces * " "):
         line = line[median_leading_spaces:]
      if line.startswith(" "):
         line = "\t" + line[1:]
      if not paragraphs or line.startswith("\t"):
         paragraphs.append([])
         paragraphs.append([])
      paragraphs[-1].append(line.rstrip())
   for paragraph in paragraphs:
      print(" ".join(paragraph))

if __name__ == "__main__":
   main()