Spaces:
Runtime error
Runtime error
File size: 4,424 Bytes
b17d312 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 |
import json
import os
import google.generativeai as genai
import gradio as gr
import pandas as pd
from gradio_pdf import PDF
from pdf2image import convert_from_path
from pypdf import PdfReader
from pathlib import Path
dir_ = Path(__file__).parent
genai.configure(api_key=os.environ.get("GOOGLE_API_KEY"))
headers=[
"DUE DATE",
"SERVICE ADDRESS",
"SERVICE PERIOD",
"ELECTRICITY USAGE (KWH)",
"ELECTRICITY SPEND ($)",
"GAS USAGE (THERMS)",
"GAS SPEND ($)",
"WATER USAGE (CCF)",
"WATER SPEND ($)",
"SEWER ($)",
"REFUSE ($)",
"STORM DRAIN ($)",
"UTILITY USERS TAX ($)",
"TOTAL CURRENT CHARGES ($)",
"TOTAL AMOUNT DUE",
]
inputs = [PDF(label="Document")]
outputs = [
gr.Dataframe(
row_count=(1, "dynamic"),
col_count=(15, "fixed"),
label="Utility",
headers=headers,
datatype=[
"str",
"str",
"str",
"str",
"str",
"str",
"str",
"str",
"str",
"str",
"str",
"str",
"str",
"str",
"str",
],
)
]
def get_content_between_curly_braces(text):
"""
This function extracts the content between the opening and closing curly braces of a string.
Args:
text: The string to extract content from.
Returns:
The extracted content as a string, or None if no curly braces are found.
"""
start_index = text.find("{")
end_index = text.rfind("}")
if start_index != -1 and end_index > start_index:
return text[start_index : end_index + 1]
else:
return None
def parse_utility_bill(filepath):
print("FOUND PDF!")
reader = PdfReader(filepath)
number_of_pages = len(reader.pages)
images = convert_from_path(filepath)
assert number_of_pages == len(images)
page = reader.pages[0]
text = page.extract_text()
image = images[0]
print("---------------------------------------------------------------")
print(f"We have the image at: ")
print(image)
print(f"Here is the text:")
print(text)
print("---------------------------------------------------------------")
model = genai.GenerativeModel(
"gemini-pro-vision",
)
promt_text = (
f""" Please extract the following JSON object from the utility bill I give. Here is the noisy OCR extractio of the page {text}. Depending on the document, it may contain values for only a few keys such as SEWER. So, you have to be extra carefull."""
+ """This JSON schema:
{'type': 'object', 'properties': { 'DUE DATE': {'type': 'string'},'SERVICE ADDRESS': {'type': 'string'},'SERVICE PERIOD': {'type': 'string'}'ELECTRICITY USAGE (KWH)': {'type': 'string'},'ELECTRICITY SPEND ($)': {'type': 'string'},'GAS USAGE (THERMS)': {'type': 'string'},'GAS SPEND ($)': {'type': 'string'},'WATER USAGE (CCF)': {'type': 'string'},'WATER SPEND ($)': {'type': 'string'},'SEWER ($)': {'type': 'string'},'REFUSE ($)': {'type': 'string'},'STORM DRAIN ($)': {'type': 'string'},'UTILITY USERS TAX ($)': {'type': 'string'},'TOTAL CURRENT CHARGES ($)': {'type': 'string'},'TOTAL AMOUNT DUE ($)': {'type': 'string'}}."""
)
print(f"PROMPT: {promt_text}")
response = model.generate_content(
[
promt_text,
image,
],
generation_config={"max_output_tokens": 2048, "temperature": 0.0},
)
json_response = get_content_between_curly_braces(response.text)
respone_dict = json.loads(json_response)
print(respone_dict)
rectified_dict = {}
for target_key in headers:
for key, value in respone_dict.items():
if key == target_key:
rectified_dict[key] = value
break
else:
rectified_dict[target_key] = None
print(rectified_dict)
example_data = [rectified_dict]
return pd.DataFrame(example_data)
gr.Interface(
fn=parse_utility_bill,
inputs=inputs,
outputs=outputs,
examples=["utl-bill-sample.pdf", "nem-2-utility-bill-sample.pdf", "Sample_Utility_Bill.pdf", "Water Bill Sample.pdf", "canada.pdf", "water.pdf"],
title="🌏⚡💧🔥PDF Utitlity Bill Parser",
).launch()
|