import sys
import os
sys.path.append(os.path.abspath("../"))

import argparse
from enum import Enum
import io
from gcloud import storage
from google.cloud import vision
from google.cloud.vision import types
from PIL import Image, ImageDraw
import os
import tempfile
from pdf2image import convert_from_path, convert_from_bytes
from request_handling_aws import get_text, get_all_text
from patient import Patient
import re
import pandas as pd
import nltk

block_markers = ['<START>', 'PATIENT NAME/ADDRESS', 'PRIMARY PLAN NAME/ADDRESS', 'SUBSCRIBER NAME/ADDRESS', 'EMPLOYER NAME/ADDRESS']
breaking_phrase = 'Advocate Illinois Masonic Medical Center'

def get_patients(text_block, block_markers,breaking_phrase):

    patient_list = []

    text_block = ['<START>'] + text_block
    blocks = {x:[] for x in block_markers}
    curr_marker = ''
    curr_patient = Patient(block_markers=block_markers)

    for line in text_block:
        nxt_ln = 0
        if breaking_phrase in line.strip():
            for key,val in blocks.items():
                print("\n\n" + key)
                print(val)
            curr_patient.process_gen_info(blocks)
            patient_list.append(curr_patient)
            curr_marker = '<START>'
            blocks = {x:[] for x in block_markers}
            curr_patient = Patient(block_markers=block_markers)

        for block_marker in block_markers:
            if nltk.edit_distance(block_marker, line.strip())<3:
                curr_marker = block_marker
                nxt_ln = 1
                break

        if nxt_ln == 1:
            continue

        blocks[curr_marker].append(line)
    for key,val in blocks.items():
        print("\n\n" + key)
        print(val)
    curr_patient.process_gen_info(blocks)
    patient_list.append(curr_patient)
    return patient_list

def compile_dataframe(patient_list):
    pat_df = pd.DataFrame()
    for patient in patient_list:
        pat_df = pd.concat([pat_df,pd.DataFrame([patient.csv_rep()], columns=patient.csv_rep().keys())],axis=0,join='outer').reset_index(drop=True)
    pat_df = pat_df.dropna(axis=1, how='all')

    dob_cols = [col for col in pat_df.columns if 'dob' in col]
    print("DOB COLS:", dob_cols)
    print(pat_df)
    for col in dob_cols:
        pat_df[col] = pat_df[col].astype(str).apply(lambda x:  x.split(" ")[0] if (x != None) else x)
    return pat_df.iloc[1:]

# full_body = get_text('gs://report-ap/test_image.jpg').full_text_annotation.text.splitlines()
def run_pipeline(full_body, form_data=None):
    print("RUNNING CAA PIPELINE........")
    block_markers = ['<START>', 'PATIENT NAME/ADDRESS', 'PRIMARY PLAN NAME/ADDRESS', 'SUBSCRIBER NAME/ADDRESS', 'EMPLOYER NAME/ADDRESS']
    breaking_phrase = 'Advocate Illinois Masonic Medical Center'
    patient_list = get_patients(full_body,block_markers,breaking_phrase)
    fin_df = compile_dataframe(patient_list)
    return fin_df

if __name__ == "__main__":
    full_body, ids = get_all_text("facesheet-ap","facesheet_caa/")
    record = []
    pat_fl = 0
    nam_nl = 0
    block_markers = ['<START>', 'PATIENT NAME/ADDRESS', 'PRIMARY PLAN NAME/ADDRESS', 'SUBSCRIBER NAME/ADDRESS', 'EMPLOYER NAME/ADDRESS']
    breaking_phrase = 'Advocate Illinois Masonic Medical Center'
    patient_list = get_patients(full_body,block_markers,breaking_phrase)
    fin_df = compile_dataframe(patient_list)
    fin_df.to_excel("./output_fin.xlsx")
    with pd.option_context('display.max_rows', None, 'display.max_columns', None):  # more options can be specified also
        print(fin_df)