Spaces:
Runtime error
Runtime error
Jayavathsan
commited on
Commit
•
d8a0a3d
1
Parent(s):
ecd571b
Upload 5 files
Browse files- .env.example +2 -0
- Invoice/invoice_1001329.pdf +0 -0
- Invoice/invoice_2001321.pdf +0 -0
- Invoice/invoice_3452334.pdf +0 -0
- utils.py +92 -0
.env.example
ADDED
@@ -0,0 +1,2 @@
|
|
|
|
|
|
|
1 |
+
REPLICATE_API_TOKEN = ""
|
2 |
+
OPENAI_API_KEY=""
|
Invoice/invoice_1001329.pdf
ADDED
Binary file (23.3 kB). View file
|
|
Invoice/invoice_2001321.pdf
ADDED
Binary file (23.3 kB). View file
|
|
Invoice/invoice_3452334.pdf
ADDED
Binary file (22.3 kB). View file
|
|
utils.py
ADDED
@@ -0,0 +1,92 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
from langchain.llms import OpenAI
|
2 |
+
from pypdf import PdfReader
|
3 |
+
from langchain.llms.openai import OpenAI
|
4 |
+
import pandas as pd
|
5 |
+
import re
|
6 |
+
import replicate
|
7 |
+
from langchain.prompts import PromptTemplate
|
8 |
+
|
9 |
+
#Extract Information from PDF file
|
10 |
+
def get_pdf_text(pdf_doc):
|
11 |
+
text = ""
|
12 |
+
pdf_reader = PdfReader(pdf_doc)
|
13 |
+
for page in pdf_reader.pages:
|
14 |
+
text += page.extract_text()
|
15 |
+
return text
|
16 |
+
|
17 |
+
|
18 |
+
|
19 |
+
#Function to extract data from text
|
20 |
+
def extracted_data(pages_data):
|
21 |
+
|
22 |
+
template = """Extract all the following values : invoice no., Description, Quantity, date,
|
23 |
+
Unit price , Amount, Total, email, phone number and address from this data: {pages}
|
24 |
+
|
25 |
+
Expected output: remove any dollar symbols {{'Invoice no.': '1001329','Description': 'Office Chair','Quantity': '2','Date': '5/4/2023','Unit price': '1100.00','Amount': '2200.00','Total': '2200.00','Email': 'Santoshvarma0988@gmail.com','Phone number': '9999999999','Address': 'Mumbai, India'}}
|
26 |
+
"""
|
27 |
+
prompt_template = PromptTemplate(input_variables=["pages"], template=template)
|
28 |
+
|
29 |
+
llm = OpenAI(temperature=.7)
|
30 |
+
full_response=llm(prompt_template.format(pages=pages_data))
|
31 |
+
|
32 |
+
|
33 |
+
#The below code will be used when we want to use LLAMA 2 model, we will use Replicate for hosting our model...
|
34 |
+
|
35 |
+
#output = replicate.run('replicate/llama-2-70b-chat:2c1608e18606fad2812020dc541930f2d0495ce32eee50074220b87300bc16e1',
|
36 |
+
#input={"prompt":prompt_template.format(pages=pages_data) ,
|
37 |
+
#"temperature":0.1, "top_p":0.9, "max_length":512, "repetition_penalty":1})
|
38 |
+
|
39 |
+
#full_response = ''
|
40 |
+
#for item in output:
|
41 |
+
#full_response += item
|
42 |
+
|
43 |
+
|
44 |
+
#print(full_response)
|
45 |
+
return full_response
|
46 |
+
|
47 |
+
|
48 |
+
# iterate over files in
|
49 |
+
# that user uploaded PDF files, one by one
|
50 |
+
def create_docs(user_pdf_list):
|
51 |
+
|
52 |
+
df = pd.DataFrame({'Invoice no.': pd.Series(dtype='str'),
|
53 |
+
'Description': pd.Series(dtype='str'),
|
54 |
+
'Quantity': pd.Series(dtype='str'),
|
55 |
+
'Date': pd.Series(dtype='str'),
|
56 |
+
'Unit price': pd.Series(dtype='str'),
|
57 |
+
'Amount': pd.Series(dtype='int'),
|
58 |
+
'Total': pd.Series(dtype='str'),
|
59 |
+
'Email': pd.Series(dtype='str'),
|
60 |
+
'Phone number': pd.Series(dtype='str'),
|
61 |
+
'Address': pd.Series(dtype='str')
|
62 |
+
})
|
63 |
+
|
64 |
+
for filename in user_pdf_list:
|
65 |
+
|
66 |
+
print(filename)
|
67 |
+
raw_data=get_pdf_text(filename)
|
68 |
+
#print(raw_data)
|
69 |
+
#print("extracted raw data")
|
70 |
+
|
71 |
+
llm_extracted_data=extracted_data(raw_data)
|
72 |
+
#print("llm extracted data")
|
73 |
+
#Adding items to our list - Adding data & its metadata
|
74 |
+
|
75 |
+
pattern = r'{(.+)}'
|
76 |
+
match = re.search(pattern, llm_extracted_data, re.DOTALL)
|
77 |
+
|
78 |
+
if match:
|
79 |
+
extracted_text = match.group(1)
|
80 |
+
# Converting the extracted text to a dictionary
|
81 |
+
data_dict = eval('{' + extracted_text + '}')
|
82 |
+
print(data_dict)
|
83 |
+
else:
|
84 |
+
print("No match found.")
|
85 |
+
|
86 |
+
|
87 |
+
df=df.append([data_dict], ignore_index=True)
|
88 |
+
print("********************DONE***************")
|
89 |
+
#df=df.append(save_to_dataframe(llm_extracted_data), ignore_index=True)
|
90 |
+
|
91 |
+
df.head()
|
92 |
+
return df
|