contract verification
Browse files- app.py +146 -0
- requirements.txt +6 -0
app.py
ADDED
@@ -0,0 +1,146 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import streamlit as st
|
2 |
+
import os
|
3 |
+
import openai
|
4 |
+
import json
|
5 |
+
import pandas as pd
|
6 |
+
from docx import Document
|
7 |
+
from concurrent.futures import ThreadPoolExecutor, as_completed
|
8 |
+
from dotenv import load_dotenv
|
9 |
+
|
10 |
+
# Load the OpenAI API key from environment variables
|
11 |
+
load_dotenv()
|
12 |
+
api_key = os.getenv("OPENAI_API_KEY")
|
13 |
+
openai.api_key = api_key
|
14 |
+
|
15 |
+
def extract_text_from_docx(docx_path):
|
16 |
+
doc = Document(docx_path)
|
17 |
+
return "\n".join([para.text for para in doc.paragraphs])
|
18 |
+
|
19 |
+
def extract_terms_from_contract(contract_text):
|
20 |
+
prompt = (
|
21 |
+
"You are an AI tasked with analyzing a contract and extracting key terms and constraints. The contract contains "
|
22 |
+
"various sections and subsections with terms related to budget constraints, types of allowable work, timelines, "
|
23 |
+
"penalties, responsibilities, and other conditions for work execution. Your job is to extract these key terms and "
|
24 |
+
"structure them in a clear JSON format, reflecting the hierarchy of sections and subsections. "
|
25 |
+
"Ensure to capture all important constraints and conditions specified in the contract text. If a section or subsection "
|
26 |
+
"contains multiple terms, list them all.\n\n"
|
27 |
+
"Contract text:\n"
|
28 |
+
f"{contract_text}\n\n"
|
29 |
+
"Provide the extracted terms in JSON format."
|
30 |
+
)
|
31 |
+
|
32 |
+
response = openai.ChatCompletion.create(
|
33 |
+
model="gpt-4",
|
34 |
+
messages=[
|
35 |
+
{"role": "system", "content": "You are an AI specialized in extracting structured data from text documents."},
|
36 |
+
{"role": "user", "content": prompt},
|
37 |
+
],
|
38 |
+
max_tokens=4096,
|
39 |
+
n=1,
|
40 |
+
stop=None,
|
41 |
+
temperature=0.1,
|
42 |
+
)
|
43 |
+
return response.choices[0].message["content"]
|
44 |
+
|
45 |
+
def analyze_task_compliance(task_description, cost_estimate, contract_terms):
|
46 |
+
prompt = (
|
47 |
+
"You are an AI tasked with analyzing a task description and its associated cost estimate for compliance with contract conditions. "
|
48 |
+
"Below are the key terms and constraints extracted from the contract, followed by a task description and its cost estimate. "
|
49 |
+
"Your job is to analyze the task description and specify if it violates any conditions from the contract. "
|
50 |
+
"If there are violations, list the reasons for each violation.\n\n"
|
51 |
+
f"Contract terms:\n{json.dumps(contract_terms, indent=4)}\n\n"
|
52 |
+
f"Task description:\n{task_description}\n"
|
53 |
+
f"Cost estimate:\n{cost_estimate}\n\n"
|
54 |
+
"Provide the compliance analysis in a clear JSON format."
|
55 |
+
)
|
56 |
+
|
57 |
+
response = openai.ChatCompletion.create(
|
58 |
+
model="gpt-4",
|
59 |
+
messages=[
|
60 |
+
{"role": "system", "content": "You are an AI specialized in analyzing text for compliance with specified conditions."},
|
61 |
+
{"role": "user", "content": prompt},
|
62 |
+
],
|
63 |
+
max_tokens=4096,
|
64 |
+
n=1,
|
65 |
+
stop=None,
|
66 |
+
temperature=0.1,
|
67 |
+
)
|
68 |
+
|
69 |
+
# Parse the response to extract structured explanations
|
70 |
+
response_content = response.choices[0].message["content"]
|
71 |
+
compliance_analysis = json.loads(response_content)
|
72 |
+
|
73 |
+
return compliance_analysis
|
74 |
+
|
75 |
+
def main():
|
76 |
+
st.title("Contract Compliance Analyzer")
|
77 |
+
|
78 |
+
# File upload buttons in the same row
|
79 |
+
col1, col2 = st.columns(2)
|
80 |
+
|
81 |
+
with col1:
|
82 |
+
docx_file = st.file_uploader("Upload Contract Document (DOCX)", type="docx")
|
83 |
+
|
84 |
+
with col2:
|
85 |
+
xlsx_file = st.file_uploader("Upload Task Descriptions (XLSX)", type="xlsx")
|
86 |
+
|
87 |
+
if docx_file and xlsx_file:
|
88 |
+
# Extract contract text and terms
|
89 |
+
contract_text = extract_text_from_docx(docx_file)
|
90 |
+
extracted_terms_json = extract_terms_from_contract(contract_text)
|
91 |
+
|
92 |
+
try:
|
93 |
+
contract_terms = json.loads(extracted_terms_json)
|
94 |
+
except json.JSONDecodeError as e:
|
95 |
+
st.error(f"JSON decoding error: {e}")
|
96 |
+
return
|
97 |
+
|
98 |
+
# Read task descriptions and cost estimates from XLSX
|
99 |
+
tasks_df = pd.read_excel(xlsx_file)
|
100 |
+
|
101 |
+
compliance_results = []
|
102 |
+
futures = []
|
103 |
+
|
104 |
+
# Use ThreadPoolExecutor to analyze tasks concurrently
|
105 |
+
with ThreadPoolExecutor(max_workers=10) as executor: # Adjust max_workers as needed
|
106 |
+
for _, row in tasks_df.iterrows():
|
107 |
+
task_description = row['Task Description']
|
108 |
+
cost_estimate = row['Amount']
|
109 |
+
futures.append(executor.submit(analyze_task_compliance, task_description, cost_estimate, contract_terms))
|
110 |
+
|
111 |
+
for future in as_completed(futures):
|
112 |
+
try:
|
113 |
+
result = future.result()
|
114 |
+
compliance_results.append(result)
|
115 |
+
except Exception as e:
|
116 |
+
st.error(f"An error occurred: {e}")
|
117 |
+
|
118 |
+
col1, col2 = st.columns(2)
|
119 |
+
|
120 |
+
with col1:
|
121 |
+
st.write("Extracted Contract Terms:")
|
122 |
+
st.json(contract_terms)
|
123 |
+
|
124 |
+
# Download button for contract terms
|
125 |
+
st.download_button(
|
126 |
+
label="Download Contract Terms",
|
127 |
+
data=json.dumps(contract_terms, indent=4),
|
128 |
+
file_name="contract_terms.json",
|
129 |
+
mime="application/json"
|
130 |
+
)
|
131 |
+
|
132 |
+
with col2:
|
133 |
+
st.write("Compliance Results:")
|
134 |
+
st.json(compliance_results)
|
135 |
+
|
136 |
+
# Download button for compliance results
|
137 |
+
compliance_results_json = json.dumps(compliance_results, indent=4)
|
138 |
+
st.download_button(
|
139 |
+
label="Download Compliance Results",
|
140 |
+
data=compliance_results_json,
|
141 |
+
file_name="compliance_results.json",
|
142 |
+
mime="application/json"
|
143 |
+
)
|
144 |
+
|
145 |
+
if __name__ == "__main__":
|
146 |
+
main()
|
requirements.txt
ADDED
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
openai==0.28
|
2 |
+
python-dotenv
|
3 |
+
flask
|
4 |
+
python-docx
|
5 |
+
pandas
|
6 |
+
streamlit
|