Spaces:
Sleeping
Sleeping
Commit ·
2fa693e
1
Parent(s): c0ab6fd
Fix flagging and possible race conditions
Browse files
app.py
CHANGED
|
@@ -1,10 +1,11 @@
|
|
| 1 |
import base64
|
| 2 |
import os
|
|
|
|
| 3 |
from io import BytesIO
|
| 4 |
from pathlib import Path
|
| 5 |
|
| 6 |
-
from langchain.schema.output_parser import OutputParserException
|
| 7 |
import gradio as gr
|
|
|
|
| 8 |
from PIL import Image
|
| 9 |
|
| 10 |
import categories
|
|
@@ -40,7 +41,6 @@ hf_writer_incorrect = gr.HuggingFaceDatasetSaver(
|
|
| 40 |
# global example_paths, current_file_path
|
| 41 |
# if current_file_path not in example_paths:
|
| 42 |
# return function(*args, **kwargs)
|
| 43 |
-
|
| 44 |
|
| 45 |
|
| 46 |
def display_file(input_file):
|
|
@@ -76,63 +76,72 @@ def clear_inputs():
|
|
| 76 |
return gr.File.update(value=None)
|
| 77 |
|
| 78 |
|
| 79 |
-
def
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 80 |
if not input_file:
|
| 81 |
gr.Error("Please upload a file to continue!")
|
| 82 |
return gr.Textbox.update()
|
| 83 |
-
print("-"*5)
|
| 84 |
-
print("New input")
|
| 85 |
# Send change to preprocessed image or to extracted text
|
| 86 |
if input_file.name.endswith(".pdf"):
|
| 87 |
text = process_pdf(Path(input_file.name), extract_only=True)
|
| 88 |
else:
|
| 89 |
text = process_image(Path(input_file.name), extract_only=True)
|
| 90 |
-
print("Extracted text")
|
| 91 |
return text
|
| 92 |
|
| 93 |
|
| 94 |
-
def
|
| 95 |
-
|
| 96 |
-
|
| 97 |
-
# gr.Info(f"Recognized category: {category}")
|
| 98 |
return category
|
| 99 |
|
| 100 |
|
| 101 |
-
def
|
| 102 |
-
|
| 103 |
-
|
| 104 |
-
|
| 105 |
-
|
| 106 |
-
gr.JSON.update(None),
|
| 107 |
-
gr.Button.update(interactive=False),
|
| 108 |
-
gr.Button.update(interactive=False),
|
| 109 |
-
)
|
| 110 |
-
else:
|
| 111 |
-
print("Updated with actual category:", category)
|
| 112 |
category = Category[category]
|
| 113 |
-
print("Parsing text from", category)
|
| 114 |
chain = categories.category_modules[category].chain
|
| 115 |
formatted_prompt = chain.prompt.format_prompt(
|
| 116 |
-
text=
|
| 117 |
format_instructions=chain.output_parser.get_format_instructions(),
|
| 118 |
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 119 |
result = chain.generate(
|
| 120 |
input_list=[
|
| 121 |
{
|
| 122 |
-
"text":
|
| 123 |
"format_instructions": chain.output_parser.get_format_instructions(),
|
| 124 |
}
|
| 125 |
]
|
| 126 |
)
|
| 127 |
-
question = f""
|
| 128 |
-
if len(formatted_prompt.messages) > 1:
|
| 129 |
-
question += f"**System:**\n{formatted_prompt.messages[0].content}"
|
| 130 |
-
question += f"\n\n**Human:**\n{formatted_prompt.messages[1].content}"
|
| 131 |
-
print("\tConstructed prompt")
|
| 132 |
answer = result.generations[0][0].text
|
| 133 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 134 |
try:
|
| 135 |
-
information = chain.output_parser.
|
| 136 |
information = information.json() if information else {}
|
| 137 |
except OutputParserException as e:
|
| 138 |
information = {
|
|
@@ -140,25 +149,22 @@ def parse_from_category(category, extracted_text):
|
|
| 140 |
"details": str(e),
|
| 141 |
"output": e.llm_output,
|
| 142 |
}
|
| 143 |
-
|
| 144 |
-
gr.Chatbot.update([[question, answer]]),
|
| 145 |
-
gr.JSON.update(information),
|
| 146 |
-
gr.Button.update(interactive=True),
|
| 147 |
-
gr.Button.update(interactive=True),
|
| 148 |
-
)
|
| 149 |
|
| 150 |
|
| 151 |
-
def
|
| 152 |
-
|
| 153 |
-
if share_result:
|
| 154 |
-
flag_method(*args, **kwargs)
|
| 155 |
|
| 156 |
-
return modified_flag_method
|
| 157 |
|
|
|
|
|
|
|
| 158 |
|
| 159 |
-
|
| 160 |
-
|
| 161 |
-
|
|
|
|
|
|
|
|
|
|
| 162 |
|
| 163 |
|
| 164 |
with gr.Blocks(title="Automatic Reimbursement Tool Demo") as page:
|
|
@@ -261,35 +267,14 @@ with gr.Blocks(title="Automatic Reimbursement Tool Demo") as page:
|
|
| 261 |
flag_irrelevant_button = gr.Button(
|
| 262 |
"Flag as irrelevant", variant="stop", interactive=True
|
| 263 |
)
|
| 264 |
-
|
| 265 |
show_intermediate.change(
|
| 266 |
show_intermediate_outputs, show_intermediate, [intermediate_outputs]
|
| 267 |
)
|
| 268 |
|
| 269 |
-
clear.click(clear_inputs, None, [input_file])
|
| 270 |
-
|
| 271 |
-
|
| 272 |
-
[
|
| 273 |
-
[extracted_text],
|
| 274 |
-
)
|
| 275 |
-
# submit_button.click(
|
| 276 |
-
# lambda input_file, category, chatbot, information: (print("File supplied, resetting") or (
|
| 277 |
-
# gr.Dropdown.update(Category.ACCOMODATION),
|
| 278 |
-
# gr.Chatbot.update(None),
|
| 279 |
-
# gr.Textbox.update(None),
|
| 280 |
-
# )) if input_file else (print("File not supplied, keeping") or print(category, chatbot, information)),
|
| 281 |
-
# [input_file, category, chatbot, information],
|
| 282 |
-
# [category, chatbot, information],
|
| 283 |
-
# )
|
| 284 |
-
extracted_text.change(
|
| 285 |
-
categorize_extracted_text,
|
| 286 |
-
[extracted_text],
|
| 287 |
-
[category],
|
| 288 |
-
)
|
| 289 |
-
category.change(
|
| 290 |
-
parse_from_category,
|
| 291 |
-
[category, extracted_text],
|
| 292 |
-
[chatbot, information, flag_incorrect_button, flag_irrelevant_button],
|
| 293 |
)
|
| 294 |
|
| 295 |
hf_writer_normal.setup(
|
|
@@ -297,11 +282,37 @@ with gr.Blocks(title="Automatic Reimbursement Tool Demo") as page:
|
|
| 297 |
flagging_dir="flagged",
|
| 298 |
)
|
| 299 |
flag_method = gr.flagging.FlagMethod(
|
| 300 |
-
hf_writer_normal, "", "", visual_feedback=
|
| 301 |
)
|
| 302 |
-
|
| 303 |
-
|
| 304 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 305 |
share_result,
|
| 306 |
input_file,
|
| 307 |
extracted_text,
|
|
@@ -310,9 +321,8 @@ with gr.Blocks(title="Automatic Reimbursement Tool Demo") as page:
|
|
| 310 |
information,
|
| 311 |
contact,
|
| 312 |
],
|
| 313 |
-
|
| 314 |
preprocess=False,
|
| 315 |
-
queue=False,
|
| 316 |
)
|
| 317 |
|
| 318 |
hf_writer_incorrect.setup(
|
|
@@ -373,5 +383,8 @@ with gr.Blocks(title="Automatic Reimbursement Tool Demo") as page:
|
|
| 373 |
queue=False,
|
| 374 |
)
|
| 375 |
|
| 376 |
-
|
|
|
|
|
|
|
|
|
|
| 377 |
page.launch(show_api=True, show_error=True, debug=True)
|
|
|
|
| 1 |
import base64
|
| 2 |
import os
|
| 3 |
+
import re
|
| 4 |
from io import BytesIO
|
| 5 |
from pathlib import Path
|
| 6 |
|
|
|
|
| 7 |
import gradio as gr
|
| 8 |
+
from langchain.schema.output_parser import OutputParserException
|
| 9 |
from PIL import Image
|
| 10 |
|
| 11 |
import categories
|
|
|
|
| 41 |
# global example_paths, current_file_path
|
| 42 |
# if current_file_path not in example_paths:
|
| 43 |
# return function(*args, **kwargs)
|
|
|
|
| 44 |
|
| 45 |
|
| 46 |
def display_file(input_file):
|
|
|
|
| 76 |
return gr.File.update(value=None)
|
| 77 |
|
| 78 |
|
| 79 |
+
def clear_outputs(input_file):
|
| 80 |
+
if input_file:
|
| 81 |
+
return None, None, None, None
|
| 82 |
+
|
| 83 |
+
|
| 84 |
+
def extract_text(input_file):
|
| 85 |
+
"""Takes the input file and updates the extracted text"""
|
| 86 |
if not input_file:
|
| 87 |
gr.Error("Please upload a file to continue!")
|
| 88 |
return gr.Textbox.update()
|
|
|
|
|
|
|
| 89 |
# Send change to preprocessed image or to extracted text
|
| 90 |
if input_file.name.endswith(".pdf"):
|
| 91 |
text = process_pdf(Path(input_file.name), extract_only=True)
|
| 92 |
else:
|
| 93 |
text = process_image(Path(input_file.name), extract_only=True)
|
|
|
|
| 94 |
return text
|
| 95 |
|
| 96 |
|
| 97 |
+
def categorize_text(text):
|
| 98 |
+
"""Takes the extracted text and updates the category"""
|
| 99 |
+
category = categories.categorize_text(text)
|
|
|
|
| 100 |
return category
|
| 101 |
|
| 102 |
|
| 103 |
+
def query(category, text):
|
| 104 |
+
"""Takes the extracted text and category and updates the chatbot in two steps:
|
| 105 |
+
1. Construct a prompt
|
| 106 |
+
2. Generate a response
|
| 107 |
+
"""
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 108 |
category = Category[category]
|
|
|
|
| 109 |
chain = categories.category_modules[category].chain
|
| 110 |
formatted_prompt = chain.prompt.format_prompt(
|
| 111 |
+
text=text,
|
| 112 |
format_instructions=chain.output_parser.get_format_instructions(),
|
| 113 |
)
|
| 114 |
+
question = f""
|
| 115 |
+
if len(formatted_prompt.messages) > 1:
|
| 116 |
+
question += f"**System:**\n{formatted_prompt.messages[0].content}"
|
| 117 |
+
question += f"\n\n**Human:**\n{formatted_prompt.messages[1].content}"
|
| 118 |
+
yield gr.Chatbot.update([[question, "Generating..."]])
|
| 119 |
+
|
| 120 |
result = chain.generate(
|
| 121 |
input_list=[
|
| 122 |
{
|
| 123 |
+
"text": text,
|
| 124 |
"format_instructions": chain.output_parser.get_format_instructions(),
|
| 125 |
}
|
| 126 |
]
|
| 127 |
)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 128 |
answer = result.generations[0][0].text
|
| 129 |
+
yield gr.Chatbot.update([[question, answer]])
|
| 130 |
+
|
| 131 |
+
|
| 132 |
+
PARSING_REGEXP = r"\*\*System:\*\*\n([\s\S]+)\n\n\*\*Human:\*\*\n([\s\S]+)"
|
| 133 |
+
|
| 134 |
+
|
| 135 |
+
def parse(category, chatbot):
|
| 136 |
+
"""Takes the chatbot prompt and response and updates the extracted information"""
|
| 137 |
+
global PARSING_REGEXP
|
| 138 |
+
|
| 139 |
+
answer = chatbot[0][1]
|
| 140 |
+
category = Category[category]
|
| 141 |
+
chain = categories.category_modules[category].chain
|
| 142 |
+
yield {"status": "Parsing response..."}
|
| 143 |
try:
|
| 144 |
+
information = chain.output_parser.parse(answer)
|
| 145 |
information = information.json() if information else {}
|
| 146 |
except OutputParserException as e:
|
| 147 |
information = {
|
|
|
|
| 149 |
"details": str(e),
|
| 150 |
"output": e.llm_output,
|
| 151 |
}
|
| 152 |
+
yield information
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 153 |
|
| 154 |
|
| 155 |
+
def activate_flags():
|
| 156 |
+
return gr.Button.update(interactive=True), gr.Button.update(interactive=True)
|
|
|
|
|
|
|
| 157 |
|
|
|
|
| 158 |
|
| 159 |
+
def deactivate_flags():
|
| 160 |
+
return gr.Button.update(interactive=False), gr.Button.update(interactive=False)
|
| 161 |
|
| 162 |
+
|
| 163 |
+
def flag_if_shared(flag_method):
|
| 164 |
+
def proxy(share_result, request: gr.Request, *args, **kwargs):
|
| 165 |
+
if share_result:
|
| 166 |
+
return flag_method(request, *args, **kwargs)
|
| 167 |
+
return proxy
|
| 168 |
|
| 169 |
|
| 170 |
with gr.Blocks(title="Automatic Reimbursement Tool Demo") as page:
|
|
|
|
| 267 |
flag_irrelevant_button = gr.Button(
|
| 268 |
"Flag as irrelevant", variant="stop", interactive=True
|
| 269 |
)
|
|
|
|
| 270 |
show_intermediate.change(
|
| 271 |
show_intermediate_outputs, show_intermediate, [intermediate_outputs]
|
| 272 |
)
|
| 273 |
|
| 274 |
+
clear.click(clear_inputs, None, [input_file]).then(
|
| 275 |
+
deactivate_flags,
|
| 276 |
+
None,
|
| 277 |
+
[flag_incorrect_button, flag_irrelevant_button],
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 278 |
)
|
| 279 |
|
| 280 |
hf_writer_normal.setup(
|
|
|
|
| 282 |
flagging_dir="flagged",
|
| 283 |
)
|
| 284 |
flag_method = gr.flagging.FlagMethod(
|
| 285 |
+
hf_writer_normal, "", "", visual_feedback=False
|
| 286 |
)
|
| 287 |
+
|
| 288 |
+
submit_button.click(
|
| 289 |
+
clear_outputs,
|
| 290 |
+
[input_file],
|
| 291 |
+
[extracted_text, category, chatbot, information],
|
| 292 |
+
).then(
|
| 293 |
+
extract_text,
|
| 294 |
+
[input_file],
|
| 295 |
+
[extracted_text],
|
| 296 |
+
).then(
|
| 297 |
+
categorize_text,
|
| 298 |
+
[extracted_text],
|
| 299 |
+
[category],
|
| 300 |
+
).then(
|
| 301 |
+
query,
|
| 302 |
+
[category, extracted_text],
|
| 303 |
+
[chatbot],
|
| 304 |
+
queue=True,
|
| 305 |
+
).then(
|
| 306 |
+
parse,
|
| 307 |
+
[category, chatbot],
|
| 308 |
+
[information],
|
| 309 |
+
).then(
|
| 310 |
+
activate_flags,
|
| 311 |
+
None,
|
| 312 |
+
[flag_incorrect_button, flag_irrelevant_button],
|
| 313 |
+
).then(
|
| 314 |
+
flag_if_shared(flag_method),
|
| 315 |
+
[
|
| 316 |
share_result,
|
| 317 |
input_file,
|
| 318 |
extracted_text,
|
|
|
|
| 321 |
information,
|
| 322 |
contact,
|
| 323 |
],
|
| 324 |
+
None,
|
| 325 |
preprocess=False,
|
|
|
|
| 326 |
)
|
| 327 |
|
| 328 |
hf_writer_incorrect.setup(
|
|
|
|
| 383 |
queue=False,
|
| 384 |
)
|
| 385 |
|
| 386 |
+
page.queue(
|
| 387 |
+
concurrency_count=1,
|
| 388 |
+
max_size=1,
|
| 389 |
+
)
|
| 390 |
page.launch(show_api=True, show_error=True, debug=True)
|