add diff
Browse files- src/app.py +61 -46
- src/suggest.py +8 -3
- src/utils.py +36 -26
src/app.py
CHANGED
|
@@ -3,6 +3,7 @@ import gradio as gr
|
|
| 3 |
from suggest import Suggest
|
| 4 |
from edit import Editor
|
| 5 |
from config import configure_logging
|
|
|
|
| 6 |
|
| 7 |
|
| 8 |
configure_logging()
|
|
@@ -10,59 +11,73 @@ configure_logging()
|
|
| 10 |
|
| 11 |
with gr.Blocks() as demo:
|
| 12 |
|
| 13 |
-
title = gr.Button("PaperGPT", interactive=
|
| 14 |
key = gr.Textbox(label="openai_key", value=os.environ.get('OPENAI_API_KEY'))
|
| 15 |
|
| 16 |
-
with gr.
|
| 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 |
# demo.launch(server_name="0.0.0.0", server_port=7653, share=True, enable_queue=True)
|
| 68 |
demo.launch(enable_queue=True)
|
|
|
|
| 3 |
from suggest import Suggest
|
| 4 |
from edit import Editor
|
| 5 |
from config import configure_logging
|
| 6 |
+
from utils import diff_texts
|
| 7 |
|
| 8 |
|
| 9 |
configure_logging()
|
|
|
|
| 11 |
|
| 12 |
with gr.Blocks() as demo:
|
| 13 |
|
| 14 |
+
title = gr.Button("PaperGPT", interactive=True)
|
| 15 |
key = gr.Textbox(label="openai_key", value=os.environ.get('OPENAI_API_KEY'))
|
| 16 |
|
| 17 |
+
with gr.Row():
|
| 18 |
+
with gr.Tab("Edit"):
|
| 19 |
+
|
| 20 |
+
handler = Editor()
|
| 21 |
+
txt_in = gr.Textbox(label="Input", lines=11, max_lines=11, value=handler.sample_content)
|
| 22 |
+
btn = gr.Button("Edit")
|
| 23 |
+
txt_out = gr.Textbox(label="Output", lines=11, max_lines=11, value="GPT will serve as your editor and modify the paragraph for you.")
|
| 24 |
+
btn.click(handler.generate, inputs=[txt_in, key], outputs=[txt_out])
|
| 25 |
+
|
| 26 |
+
with gr.Tab("Suggest"):
|
| 27 |
+
|
| 28 |
+
max_ideas = 5
|
| 29 |
+
handler = Suggest(max_ideas)
|
| 30 |
|
| 31 |
+
def select(name: str):
|
| 32 |
+
for i in handler.idea_list:
|
| 33 |
+
if i['title'] == name:
|
| 34 |
+
return [
|
| 35 |
+
gr.Textbox.update(value=i["thought"], label="thought", visible=True),
|
| 36 |
+
gr.Textbox.update(value=i["action"], label="action", visible=True),
|
| 37 |
+
gr.Textbox.update(value=i["original"], label="original", visible=True, max_lines=5, lines=5),
|
| 38 |
+
gr.Textbox.update(value=i["improved"], label="improved", visible=True, max_lines=5, lines=5),
|
| 39 |
+
gr.HighlightedText.update(value=diff_texts(i["original"], i["improved"]), visible=True)
|
| 40 |
+
]
|
| 41 |
|
| 42 |
+
with gr.Row().style(equal_height=True):
|
| 43 |
+
with gr.Column(scale=0.95):
|
| 44 |
+
txt_in = gr.Textbox(label="Input", lines=11, max_lines=11, value=handler.sample_content[2048+2048+256-45:])
|
| 45 |
+
with gr.Column(scale=0.05):
|
| 46 |
+
upload = gr.File(file_count="single", file_types=["tex", ".pdf"])
|
| 47 |
+
btn = gr.Button("Analyze")
|
| 48 |
+
upload.change(handler.read_file, inputs=upload, outputs=txt_in)
|
| 49 |
|
| 50 |
+
textboxes = []
|
| 51 |
+
sug = gr.Textbox("GPT will give suggestions and help you improve the paper quality.", interactive=False, show_label=False, lines=11).style(text_align="center")
|
| 52 |
+
with gr.Row():
|
| 53 |
+
with gr.Column(scale=0.4):
|
| 54 |
+
for i in range(max_ideas):
|
| 55 |
+
t = gr.Button("", visible=False)
|
| 56 |
+
textboxes.append(t)
|
| 57 |
+
with gr.Column(scale=0.6):
|
| 58 |
+
thought = gr.Textbox(label="thought", visible=False, interactive=False)
|
| 59 |
+
action = gr.Textbox(label="action", visible=False, interactive=False)
|
| 60 |
+
original = gr.Textbox(label="original", visible=False, max_lines=5, lines=5, interactive=False)
|
| 61 |
+
improved = gr.Textbox(label="improved", visible=False, max_lines=5, lines=5, interactive=False)
|
| 62 |
+
diff = gr.HighlightedText(
|
| 63 |
+
label="Diff",
|
| 64 |
+
combine_adjacent=True,
|
| 65 |
+
show_legend=True,
|
| 66 |
+
visible=False,
|
| 67 |
+
max_lines=5,
|
| 68 |
+
lines=5,
|
| 69 |
+
interactive=False
|
| 70 |
+
).style(color_map={"+": "green", "-": "red"})
|
| 71 |
|
| 72 |
+
btn.click(handler.generate, inputs=[txt_in, key], outputs=[sug, btn, thought, action, original, improved] + textboxes)
|
| 73 |
+
for i in textboxes:
|
| 74 |
+
i.click(select, inputs=[i], outputs=[thought, action, original, improved, diff])
|
| 75 |
|
| 76 |
+
with gr.Row():
|
| 77 |
+
with gr.Tab("Issue"):
|
| 78 |
+
gr.Textbox(show_label=False, value="https://github.com/j40903272/PaperGPT/issues", interactive=False)
|
| 79 |
+
with gr.Tab("Author"):
|
| 80 |
+
gr.JSON(show_label=False, value={'author': 'YDTsai', 'email': 'bb04902103@gmail.com', 'source': 'https://github.com/j40903272/PaperGPT'})
|
| 81 |
|
| 82 |
# demo.launch(server_name="0.0.0.0", server_port=7653, share=True, enable_queue=True)
|
| 83 |
demo.launch(enable_queue=True)
|
src/suggest.py
CHANGED
|
@@ -52,6 +52,7 @@ class Suggest():
|
|
| 52 |
self.max_ideas = max_ideas
|
| 53 |
self.encoder = tiktoken.encoding_for_model(model)
|
| 54 |
self.model = model
|
|
|
|
| 55 |
with open("./sample/sample.tex", "r") as f:
|
| 56 |
self.sample_content = f.read()
|
| 57 |
|
|
@@ -125,7 +126,11 @@ class Suggest():
|
|
| 125 |
if len(ideas) >= self.max_ideas:
|
| 126 |
break
|
| 127 |
else:
|
| 128 |
-
raise gr.Error(idea)
|
|
|
|
|
|
|
|
|
|
|
|
|
| 129 |
|
| 130 |
logging.info('complete analysis')
|
| 131 |
return ideas
|
|
@@ -149,8 +154,8 @@ class Suggest():
|
|
| 149 |
raise gr.Error("Please provide openai key !")
|
| 150 |
|
| 151 |
try:
|
| 152 |
-
global idea_list
|
| 153 |
idea_list = self.analyze(txt, openai_key, progress)
|
|
|
|
| 154 |
k = min(len(idea_list), self.max_ideas)
|
| 155 |
|
| 156 |
idea_buttons = [
|
|
@@ -165,7 +170,7 @@ class Suggest():
|
|
| 165 |
gr.Textbox.update(value="", label="thought", visible=True),
|
| 166 |
gr.Textbox.update(value="", label="action", visible=True),
|
| 167 |
gr.Textbox.update(value="", label="original", visible=True, max_lines=5, lines=5),
|
| 168 |
-
gr.Textbox.update(value="", label="improved", visible=True, max_lines=5, lines=5)
|
| 169 |
]
|
| 170 |
|
| 171 |
return [
|
|
|
|
| 52 |
self.max_ideas = max_ideas
|
| 53 |
self.encoder = tiktoken.encoding_for_model(model)
|
| 54 |
self.model = model
|
| 55 |
+
self.idea_list = []
|
| 56 |
with open("./sample/sample.tex", "r") as f:
|
| 57 |
self.sample_content = f.read()
|
| 58 |
|
|
|
|
| 126 |
if len(ideas) >= self.max_ideas:
|
| 127 |
break
|
| 128 |
else:
|
| 129 |
+
# raise gr.Error(idea)
|
| 130 |
+
continue
|
| 131 |
+
|
| 132 |
+
if not ideas:
|
| 133 |
+
raise gr.Error('No suggestions generated.')
|
| 134 |
|
| 135 |
logging.info('complete analysis')
|
| 136 |
return ideas
|
|
|
|
| 154 |
raise gr.Error("Please provide openai key !")
|
| 155 |
|
| 156 |
try:
|
|
|
|
| 157 |
idea_list = self.analyze(txt, openai_key, progress)
|
| 158 |
+
self.idea_list = idea_list
|
| 159 |
k = min(len(idea_list), self.max_ideas)
|
| 160 |
|
| 161 |
idea_buttons = [
|
|
|
|
| 170 |
gr.Textbox.update(value="", label="thought", visible=True),
|
| 171 |
gr.Textbox.update(value="", label="action", visible=True),
|
| 172 |
gr.Textbox.update(value="", label="original", visible=True, max_lines=5, lines=5),
|
| 173 |
+
gr.Textbox.update(value="", label="improved", visible=True, max_lines=5, lines=5),
|
| 174 |
]
|
| 175 |
|
| 176 |
return [
|
src/utils.py
CHANGED
|
@@ -2,6 +2,7 @@ import json
|
|
| 2 |
import requests
|
| 3 |
import tiktoken
|
| 4 |
import logging
|
|
|
|
| 5 |
|
| 6 |
|
| 7 |
def json_validator(
|
|
@@ -18,33 +19,34 @@ def json_validator(
|
|
| 18 |
return json.loads(text)
|
| 19 |
except Exception:
|
| 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 |
return text
|
| 50 |
|
|
@@ -77,3 +79,11 @@ def fetch_chat(
|
|
| 77 |
return result
|
| 78 |
|
| 79 |
return response.json()["error"]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 2 |
import requests
|
| 3 |
import tiktoken
|
| 4 |
import logging
|
| 5 |
+
from difflib import Differ
|
| 6 |
|
| 7 |
|
| 8 |
def json_validator(
|
|
|
|
| 19 |
return json.loads(text)
|
| 20 |
except Exception:
|
| 21 |
|
| 22 |
+
|
| 23 |
+
prompt = f"Modify the following into a valid json format:\n{text}"
|
| 24 |
+
prompt_token_length = len(encoder.encode(prompt))
|
| 25 |
|
| 26 |
+
data = {
|
| 27 |
+
"model": model,
|
| 28 |
+
"prompt": prompt,
|
| 29 |
+
"max_tokens": 4097 - prompt_token_length - 64
|
| 30 |
+
}
|
| 31 |
+
headers = {
|
| 32 |
+
"Content-Type": "application/json",
|
| 33 |
+
"Authorization": f"Bearer {openai_key}"
|
| 34 |
+
}
|
| 35 |
+
for _ in range(retry):
|
| 36 |
+
response = requests.post(
|
| 37 |
+
'https://api.openai.com/v1/completions',
|
| 38 |
+
json=data,
|
| 39 |
+
headers=headers,
|
| 40 |
+
timeout=300
|
| 41 |
+
)
|
| 42 |
+
if response.status_code != 200:
|
| 43 |
+
logging.warning(f'fetch openai chat retry: {response.text}')
|
| 44 |
+
continue
|
| 45 |
+
text = response.json()['choices'][0]['text']
|
| 46 |
+
break
|
| 47 |
+
|
| 48 |
+
if response.status_code != 200:
|
| 49 |
+
raise Exception(response.json()['error'])
|
| 50 |
|
| 51 |
return text
|
| 52 |
|
|
|
|
| 79 |
return result
|
| 80 |
|
| 81 |
return response.json()["error"]
|
| 82 |
+
|
| 83 |
+
|
| 84 |
+
def diff_texts(text1: str, text2: str) -> list:
|
| 85 |
+
d = Differ()
|
| 86 |
+
return [
|
| 87 |
+
(token[2:], token[0] if token[0] != " " else None)
|
| 88 |
+
for token in d.compare(text1, text2)
|
| 89 |
+
]
|