File size: 10,165 Bytes
02dad28 d870c6e 02dad28 d870c6e 02dad28 8b0e4a2 02dad28 d870c6e 02dad28 8b0e4a2 02dad28 bd6c5bf d870c6e 02dad28 d870c6e 02dad28 bd6c5bf d870c6e 02dad28 115cf1f 02dad28 d870c6e 02dad28 8b0e4a2 02dad28 8b0e4a2 02dad28 8b0e4a2 02dad28 d870c6e 02dad28 bc54de9 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 |
import os
import openai
import gradio as gr
import requests
import datetime
from io import BytesIO
from google.api_core.client_options import ClientOptions
from google.cloud import documentai_v1 as documentai
import json
from google.cloud import vision
import time
from settings import char_remove, gpt_model, RPFAAP2, RPFAAP1, project_id, project_location, processor_id
from tqdm import tqdm
import logging
logging.basicConfig(filename='app.log', level=logging.INFO, format='%(asctime)s - %(levelname)s - %(message)s')
def chat_gpt_image(content, context):
openai.api_key = os.environ['GPT_API_KEY']
prompt = "You are an expert at identifying OCR errors and correcting them with the help of context, intuition and logic."
document = "The following text was scanned using OCR, your goal is to return a corrected version of the text"
prefix = "Additionally"
if context == "":
sequence = (document, content)
else:
sequence_1 = (prefix, context)
additional = (" ".join(sequence_1))
sequence = (additional, content)
final_content = (" ".join(sequence))
logging.info(final_content)
completion = openai.ChatCompletion.create(
model=gpt_model,
user="1",
temperature="0.5",
messages=[
{"role": "system", "content": prompt},
{"role": "user", "content": final_content}
]
)
logging.info(completion.choices[0].message.content)
return(completion.choices[0].message.content)
def remove_na(string):
for char in char_remove:
string = string.replace(char, "")
return string
def chat_gpt_document(content, document_type, context):
openai.api_key = os.environ['GPT_API_KEY']
prompt = "You are an expert at identifying OCR errors and correcting them with the help of context, intuition and logic."
document_prefix = "The following text was scanned using OCR, your goal is to extract the important entities from the text and correct them with the help of the restrictions placed in the desired format. Remember to not make any changes on the labels of the desired format, simply extract the text, correct it and return only the desired format. Text:"
additional_prefix = "Additionally the text"
content_info = content[0]
content_name = content[1]
if document_type == "RPFAA Building P1":
document = "RPFAAP1.json"
desired_format = RPFAAP1
elif document_type == "RPFAA Building P2":
document = "RPFAAP2.json"
desired_format = RPFAAP2
else:
document = ""
desired_format = ""
if context == "":
sequence_1 = (document_prefix, content_info, desired_format)
else:
sequence_1 = (document_prefix, content_info, desired_format, additional_prefix, context)
content_1 = (" ".join(sequence_1))
logging.info(content_1)
completion_1 = openai.ChatCompletion.create(
model=gpt_model,
user="1",
temperature="0.1",
messages=[
{"role": "system", "content": prompt},
{"role": "user", "content": content_1}
]
)
logging.info(completion_1.choices[0].message.content)
input_string = remove_na(completion_1.choices[0].message.content)
with open(document) as f:
property_info = json.load(f)
#Adds the name of the file
property_info["File Name"] = content_name
#Fills in the information
for line in input_string.split('\n'):
if ':' in line:
key, value = line.split(':', 1)
key = key.strip()
for category in property_info:
if key in property_info[category]:
property_info[category][key] = value.strip()
break
else:
if key == "File Name":
property_info[key] = value.strip()
return json.dumps(property_info, indent=4)
def detect_image(content, lang):
credentials = json.loads(os.environ['CREDENTIALS'])
temp_file_path = 'temp_credentials.json'
with open(temp_file_path, 'w') as file:
json.dump(credentials, file)
os.environ['GOOGLE_APPLICATION_CREDENTIALS']=r'temp_credentials.json'
client = vision.ImageAnnotatorClient()
buffer = BytesIO()
content.save(buffer, format="PNG")
content = buffer.getvalue()
if lang == "Filpino":
hints = "tl"
else:
hints = "en"
image = vision.Image(content=content)
response = client.document_text_detection(image=image, image_context={"language_hints": [hints]})
if response.error.message:
raise Exception(
'{}\nFor more info on error messages, check: '
'https://cloud.google.com/apis/design/errors'.format(
response.error.message))
os.remove(temp_file_path)
return(response.full_text_annotation.text)
def detect_document(content):
credentials = json.loads(os.environ['CREDENTIALS'])
temp_file_path = 'temp_credentials.json'
with open(temp_file_path, 'w') as file:
json.dump(credentials, file)
os.environ['GOOGLE_APPLICATION_CREDENTIALS']=r'temp_credentials.json'
PROJECT_ID = project_id
LOCATION = project_location # Format is 'us' or 'eu'
PROCESSOR_ID = processor_id # Create processor in Cloud Console
content_extension = content.name.split(".")[-1]
if content_extension.upper() == "TIFF":
MIME_TYPE = "image/tiff"
elif content_extension.upper() =="PDF":
MIME_TYPE = "application/pdf"
elif content_extension.upper() =="PNG":
MIME_TYPE = "image/png"
elif content_extension.upper() =="JPG":
MIME_TYPE = "image/jpg"
else:
return("Please upload a valid MIME type")
docai_client = documentai.DocumentProcessorServiceClient(
client_options=ClientOptions(api_endpoint=f"{LOCATION}-documentai.googleapis.com")
)
RESOURCE_NAME = docai_client.processor_path(PROJECT_ID, LOCATION, PROCESSOR_ID)
with open(content.name, "rb") as image:
image_content = image.read()
raw_document = documentai.RawDocument(content=image_content, mime_type=MIME_TYPE)
request = documentai.ProcessRequest(name=RESOURCE_NAME, raw_document=raw_document)
result = docai_client.process_document(request=request)
document_object = result.document
name = content.name.split('\\')[-1]
name = name.split("/")[-1]
name = name.split('.')[0]
os.remove(temp_file_path)
return(document_object.text, name)
def image(content, lang, context):
return chat_gpt_image(detect_image(content, lang), context)
def document(content, document_type, context):
return chat_gpt_document(detect_document(content),document_type,context)
def batch_document(content, document_type, context, progress=gr.Progress()):
progress(0, desc="Starting")
retries = 5
timeout = 5
i = 0
j = 0
combined_data = []
for x in progress.tqdm(content, desc="Processing"):
while True:
try:
data = json.loads(chat_gpt_document(detect_document(x),document_type,context))
combined_data.append(data)
break
except Exception as e:
logging.error(f'Retry {i+1} failed: ', e)
if i < retries - 1:
logging.error(f'Retrying in {timeout} seconds...')
time.sleep(timeout)
i += 1
# except openai.error.RateLimitError:
# logging.error(f'Retry {j+1} failed: openai.error.RateLimitError')
# if j < retries - 1:
# logging.error(f'Retrying in {timeout} seconds...')
# time.sleep(timeout)
# j += 1
logging.info(combined_data)
return save_json(combined_data, document_type)
def save_json(text, filename):
filename = filename+".json"
with open(filename, "w", encoding='utf-8') as outfile:
json.dump(text, outfile)
return filename
with gr.Blocks(title="Ottico OCR", css=".markdown {text-align: center;}", theme='freddyaboulton/dracula_revamped') as app:
gr.Markdown("""# Ottico OCR
Attach Images or Files below and convert them to Text.""", elem_classes="markdown")
with gr.Tab("Scan Image"):
with gr.Row():
with gr.Column():
image_input = [gr.Image(type="pil"),
gr.Radio(["English", "Filipino"], label="Language", info="What is the document language? (Optional)"),
gr.Textbox(label="What kind of Image is this? (Optional)", placeholder="This is an image of an Official Reciept")]
image_output = gr.Textbox(label="Result")
image_button = gr.Button("Scan")
with gr.Tab("Scan Document"):
with gr.Row():
with gr.Column():
document_input = [gr.File(file_types=["pdf","tiff","image","text"]),
gr.Dropdown(["RPFAA Building P1", "RPFAA Building P2"], label="File Type", info="What type of document is this?"),
gr.Textbox(label="Any additional information? (Optional)", placeholder="This is document is an Official Reciept")]
document_output = gr.Textbox(label="Result")
document_button = gr.Button("Scan")
with gr.Tab("Batch Scan"):
with gr.Row():
with gr.Column():
batch_document_input = [gr.File(file_types=["pdf","tiff","image","text"], file_count="multiple"),
gr.Dropdown(["RPFAA Building P1", "RPFAA Building P2"], label="File Type", info="What type of document is this?"),
gr.Textbox(label="Any additional information? (Optional)", placeholder="This is document is an Official Reciept")]
batch_document_output = gr.File(label="Result")
batch_document_button = gr.Button("Scan")
image_button.click(image, inputs=image_input, outputs=image_output)
document_button.click(document, inputs=document_input, outputs=document_output)
batch_document_button.click(batch_document, inputs=batch_document_input, outputs=batch_document_output)
app.queue()
app.launch(auth=("username", "password"))
|