Update app.py
Browse files
app.py
CHANGED
@@ -3,7 +3,7 @@ import os
|
|
3 |
import pandas as pd
|
4 |
import json
|
5 |
import gradio as gr
|
6 |
-
from typing import List, Tuple,
|
7 |
import hashlib
|
8 |
import shutil
|
9 |
import re
|
@@ -42,20 +42,17 @@ def clean_response(text: str) -> str:
|
|
42 |
def estimate_tokens(text: str) -> int:
|
43 |
return len(text) // 3.5 + 1
|
44 |
|
45 |
-
def extract_text_from_excel(
|
46 |
all_text = []
|
47 |
-
|
48 |
-
xls = pd.ExcelFile(file_obj)
|
49 |
-
except Exception as e:
|
50 |
-
raise ValueError(f"β Error reading Excel file: {e}")
|
51 |
for sheet_name in xls.sheet_names:
|
52 |
df = xls.parse(sheet_name).astype(str).fillna("")
|
53 |
-
rows = df.apply(lambda row: " | ".join(
|
54 |
-
sheet_text = [f"[{sheet_name}] {line}" for line in rows
|
55 |
all_text.extend(sheet_text)
|
56 |
return "\n".join(all_text)
|
57 |
|
58 |
-
def split_text_into_chunks(text: str, max_tokens: int = MAX_CHUNK_TOKENS
|
59 |
effective_max = max_tokens - PROMPT_OVERHEAD
|
60 |
lines, chunks, curr_chunk, curr_tokens = text.split("\n"), [], [], 0
|
61 |
for line in lines:
|
@@ -63,13 +60,11 @@ def split_text_into_chunks(text: str, max_tokens: int = MAX_CHUNK_TOKENS, max_ch
|
|
63 |
if curr_tokens + t > effective_max:
|
64 |
if curr_chunk:
|
65 |
chunks.append("\n".join(curr_chunk))
|
66 |
-
if len(chunks) >= max_chunks:
|
67 |
-
break
|
68 |
curr_chunk, curr_tokens = [line], t
|
69 |
else:
|
70 |
curr_chunk.append(line)
|
71 |
curr_tokens += t
|
72 |
-
if curr_chunk
|
73 |
chunks.append("\n".join(curr_chunk))
|
74 |
return chunks
|
75 |
|
@@ -92,48 +87,14 @@ Analyze the following clinical notes and provide a detailed, concise summary foc
|
|
92 |
Respond in well-structured bullet points with medical reasoning.
|
93 |
"""
|
94 |
|
95 |
-
def validate_tool_file(file_path):
|
96 |
-
try:
|
97 |
-
with open(file_path, 'r') as f:
|
98 |
-
data = json.load(f)
|
99 |
-
if isinstance(data, list):
|
100 |
-
assert all(isinstance(t, dict) and "name" in t for t in data), "Invalid list format"
|
101 |
-
elif isinstance(data, dict):
|
102 |
-
assert "tools" in data and isinstance(data["tools"], list), "'tools' field missing or invalid"
|
103 |
-
assert all(isinstance(t, dict) and "name" in t for t in data["tools"]), "Invalid item in 'tools'"
|
104 |
-
else:
|
105 |
-
raise ValueError("Unexpected structure")
|
106 |
-
return True
|
107 |
-
except Exception as e:
|
108 |
-
print(f"β Tool validation failed for {file_path}: {e}")
|
109 |
-
return False
|
110 |
-
|
111 |
def init_agent():
|
112 |
-
|
113 |
-
|
114 |
-
|
115 |
-
"special_tools": "/home/user/.pyenv/versions/3.10.17/lib/python3.10/site-packages/tooluniverse/data/special_tools.json",
|
116 |
-
"monarch": "/home/user/.pyenv/versions/3.10.17/lib/python3.10/site-packages/tooluniverse/data/monarch_tools.json",
|
117 |
-
"new_tool": os.path.join(tool_cache_dir, "new_tool.json"),
|
118 |
-
}
|
119 |
-
|
120 |
-
if not os.path.exists(all_tool_paths["new_tool"]):
|
121 |
-
shutil.copy(os.path.abspath("data/new_tool.json"), all_tool_paths["new_tool"])
|
122 |
-
|
123 |
-
valid_tool_paths = {}
|
124 |
-
for key, path in all_tool_paths.items():
|
125 |
-
if validate_tool_file(path):
|
126 |
-
valid_tool_paths[key] = path
|
127 |
-
else:
|
128 |
-
print(f"β οΈ Skipping invalid tool file: {path}")
|
129 |
-
|
130 |
-
if not valid_tool_paths:
|
131 |
-
raise RuntimeError("β No valid tool files found to load into TxAgent.")
|
132 |
-
|
133 |
agent = TxAgent(
|
134 |
model_name="mims-harvard/TxAgent-T1-Llama-3.1-8B",
|
135 |
rag_model_name="mims-harvard/ToolRAG-T1-GTE-Qwen2-1.5B",
|
136 |
-
tool_files_dict=
|
137 |
force_finish=True,
|
138 |
enable_checker=True,
|
139 |
step_rag_num=4,
|
@@ -142,111 +103,126 @@ def init_agent():
|
|
142 |
agent.init_model()
|
143 |
return agent
|
144 |
|
145 |
-
def
|
146 |
-
|
147 |
-
|
148 |
-
|
149 |
-
yield "β Please upload a valid Excel file.", None, ""
|
150 |
-
return
|
151 |
|
152 |
-
|
153 |
-
|
154 |
-
|
155 |
-
|
156 |
-
else:
|
157 |
-
raise ValueError("β Invalid or missing file.")
|
158 |
-
|
159 |
-
chunks = split_text_into_chunks(text)
|
160 |
-
|
161 |
-
for i, chunk in enumerate(chunks):
|
162 |
-
prompt = build_prompt_from_text(chunk)
|
163 |
-
partial = ""
|
164 |
-
for res in agent.run_gradio_chat(
|
165 |
-
message=prompt, history=[], temperature=0.2,
|
166 |
-
max_new_tokens=MAX_NEW_TOKENS, max_token=MAX_MODEL_TOKENS,
|
167 |
-
call_agent=False, conversation=[]
|
168 |
-
):
|
169 |
-
if isinstance(res, str):
|
170 |
-
partial += res
|
171 |
-
elif hasattr(res, "content"):
|
172 |
-
partial += res.content
|
173 |
-
cleaned = clean_response(partial)
|
174 |
-
accumulated_text += f"\n\nπ **Chunk {i+1}**:\n{cleaned}"
|
175 |
-
yield accumulated_text, None, ""
|
176 |
|
177 |
-
|
178 |
-
|
179 |
-
|
180 |
-
|
181 |
-
max_new_tokens=MAX_NEW_TOKENS, max_token=MAX_MODEL_TOKENS,
|
182 |
-
call_agent=False, conversation=[]
|
183 |
-
):
|
184 |
if isinstance(res, str):
|
185 |
-
|
186 |
elif hasattr(res, "content"):
|
187 |
-
|
188 |
-
|
189 |
-
|
190 |
-
|
191 |
-
|
192 |
-
|
193 |
-
|
194 |
-
|
195 |
-
|
196 |
-
|
197 |
-
|
198 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
199 |
|
200 |
def create_ui(agent):
|
201 |
with gr.Blocks(css="""
|
202 |
-
body {
|
203 |
-
|
204 |
-
color: #
|
|
|
205 |
font-family: 'Inter', sans-serif;
|
206 |
-
margin: 0;
|
207 |
-
padding: 0;
|
208 |
}
|
209 |
-
.
|
210 |
-
|
211 |
-
|
212 |
-
|
213 |
-
|
214 |
-
|
|
|
|
|
|
|
|
|
215 |
}
|
216 |
-
.
|
217 |
-
background-color: #
|
|
|
218 |
border-radius: 12px;
|
219 |
-
|
220 |
-
min-height: 600px;
|
221 |
-
overflow-y: auto;
|
222 |
-
border: 1px solid #2c3344;
|
223 |
}
|
224 |
-
.
|
225 |
-
background:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
226 |
color: white;
|
227 |
-
font-weight: 500;
|
228 |
-
border: none;
|
229 |
-
padding: 10px 20px;
|
230 |
border-radius: 8px;
|
231 |
-
|
232 |
}
|
233 |
-
.gr-button:hover {
|
234 |
-
background:
|
235 |
}
|
236 |
""") as demo:
|
237 |
-
gr.Markdown("""
|
238 |
-
|
239 |
-
|
240 |
-
|
241 |
-
|
242 |
-
|
243 |
-
|
244 |
-
|
245 |
-
|
246 |
-
|
247 |
-
|
248 |
-
|
249 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
250 |
|
251 |
return demo
|
252 |
|
@@ -254,7 +230,7 @@ if __name__ == "__main__":
|
|
254 |
try:
|
255 |
agent = init_agent()
|
256 |
demo = create_ui(agent)
|
257 |
-
demo.launch(server_name="0.0.0.0", server_port=7860, allowed_paths=["/data/hf_cache/reports"], share=
|
258 |
except Exception as e:
|
259 |
print(f"Error: {str(e)}")
|
260 |
sys.exit(1)
|
|
|
3 |
import pandas as pd
|
4 |
import json
|
5 |
import gradio as gr
|
6 |
+
from typing import List, Tuple, Dict, Any, Union
|
7 |
import hashlib
|
8 |
import shutil
|
9 |
import re
|
|
|
42 |
def estimate_tokens(text: str) -> int:
|
43 |
return len(text) // 3.5 + 1
|
44 |
|
45 |
+
def extract_text_from_excel(file_path: str) -> str:
|
46 |
all_text = []
|
47 |
+
xls = pd.ExcelFile(file_path)
|
|
|
|
|
|
|
48 |
for sheet_name in xls.sheet_names:
|
49 |
df = xls.parse(sheet_name).astype(str).fillna("")
|
50 |
+
rows = df.apply(lambda row: " | ".join(row), axis=1)
|
51 |
+
sheet_text = [f"[{sheet_name}] {line}" for line in rows]
|
52 |
all_text.extend(sheet_text)
|
53 |
return "\n".join(all_text)
|
54 |
|
55 |
+
def split_text_into_chunks(text: str, max_tokens: int = MAX_CHUNK_TOKENS) -> List[str]:
|
56 |
effective_max = max_tokens - PROMPT_OVERHEAD
|
57 |
lines, chunks, curr_chunk, curr_tokens = text.split("\n"), [], [], 0
|
58 |
for line in lines:
|
|
|
60 |
if curr_tokens + t > effective_max:
|
61 |
if curr_chunk:
|
62 |
chunks.append("\n".join(curr_chunk))
|
|
|
|
|
63 |
curr_chunk, curr_tokens = [line], t
|
64 |
else:
|
65 |
curr_chunk.append(line)
|
66 |
curr_tokens += t
|
67 |
+
if curr_chunk:
|
68 |
chunks.append("\n".join(curr_chunk))
|
69 |
return chunks
|
70 |
|
|
|
87 |
Respond in well-structured bullet points with medical reasoning.
|
88 |
"""
|
89 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
90 |
def init_agent():
|
91 |
+
tool_path = os.path.join(tool_cache_dir, "new_tool.json")
|
92 |
+
if not os.path.exists(tool_path):
|
93 |
+
shutil.copy(os.path.abspath("data/new_tool.json"), tool_path)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
94 |
agent = TxAgent(
|
95 |
model_name="mims-harvard/TxAgent-T1-Llama-3.1-8B",
|
96 |
rag_model_name="mims-harvard/ToolRAG-T1-GTE-Qwen2-1.5B",
|
97 |
+
tool_files_dict={"new_tool": tool_path},
|
98 |
force_finish=True,
|
99 |
enable_checker=True,
|
100 |
step_rag_num=4,
|
|
|
103 |
agent.init_model()
|
104 |
return agent
|
105 |
|
106 |
+
def process_final_report(agent, file, chatbot_state: List[Dict[str, str]]) -> Tuple[List[Dict[str, str]], Union[str, None]]:
|
107 |
+
messages = chatbot_state if chatbot_state else []
|
108 |
+
if file is None or not hasattr(file, "name"):
|
109 |
+
return messages + [{"role": "assistant", "content": "β Please upload a valid Excel file."}], None
|
|
|
|
|
110 |
|
111 |
+
messages.append({"role": "user", "content": f"Processing Excel file: {os.path.basename(file.name)}"})
|
112 |
+
text = extract_text_from_excel(file.name)
|
113 |
+
chunks = split_text_into_chunks(text)
|
114 |
+
chunk_responses = [None] * len(chunks)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
115 |
|
116 |
+
def analyze_chunk(i, chunk):
|
117 |
+
prompt = build_prompt_from_text(chunk)
|
118 |
+
response = ""
|
119 |
+
for res in agent.run_gradio_chat(message=prompt, history=[], temperature=0.2, max_new_tokens=MAX_NEW_TOKENS, max_token=MAX_MODEL_TOKENS, call_agent=False, conversation=[]):
|
|
|
|
|
|
|
120 |
if isinstance(res, str):
|
121 |
+
response += res
|
122 |
elif hasattr(res, "content"):
|
123 |
+
response += res.content
|
124 |
+
elif isinstance(res, list):
|
125 |
+
for r in res:
|
126 |
+
if hasattr(r, "content"):
|
127 |
+
response += r.content
|
128 |
+
return i, clean_response(response)
|
129 |
+
|
130 |
+
with ThreadPoolExecutor(max_workers=1) as executor:
|
131 |
+
futures = [executor.submit(analyze_chunk, i, c) for i, c in enumerate(chunks)]
|
132 |
+
for f in as_completed(futures):
|
133 |
+
i, result = f.result()
|
134 |
+
chunk_responses[i] = result
|
135 |
+
|
136 |
+
valid = [r for r in chunk_responses if r and not r.startswith("β")]
|
137 |
+
if not valid:
|
138 |
+
return messages + [{"role": "assistant", "content": "β No valid chunk results."}], None
|
139 |
+
|
140 |
+
summary_prompt = f"Summarize this analysis in a final structured report:\n\n" + "\n\n".join(valid)
|
141 |
+
messages.append({"role": "assistant", "content": "π Generating final report..."})
|
142 |
+
|
143 |
+
final_report = ""
|
144 |
+
for res in agent.run_gradio_chat(message=summary_prompt, history=[], temperature=0.2, max_new_tokens=MAX_NEW_TOKENS, max_token=MAX_MODEL_TOKENS, call_agent=False, conversation=[]):
|
145 |
+
if isinstance(res, str):
|
146 |
+
final_report += res
|
147 |
+
elif hasattr(res, "content"):
|
148 |
+
final_report += res.content
|
149 |
+
|
150 |
+
cleaned = clean_response(final_report)
|
151 |
+
report_path = os.path.join(report_dir, f"report_{datetime.now().strftime('%Y%m%d_%H%M%S')}.md")
|
152 |
+
with open(report_path, 'w') as f:
|
153 |
+
f.write(f"# π§ Final Patient Report\n\n{cleaned}")
|
154 |
+
|
155 |
+
messages.append({"role": "assistant", "content": f"π Final Report:\n\n{cleaned}"})
|
156 |
+
messages.append({"role": "assistant", "content": f"β
Report generated and saved: {os.path.basename(report_path)}"})
|
157 |
+
return messages, report_path
|
158 |
|
159 |
def create_ui(agent):
|
160 |
with gr.Blocks(css="""
|
161 |
+
html, body, .gradio-container {
|
162 |
+
height: 100vh;
|
163 |
+
background-color: #111827;
|
164 |
+
color: #e5e7eb;
|
165 |
font-family: 'Inter', sans-serif;
|
|
|
|
|
166 |
}
|
167 |
+
.message-avatar {
|
168 |
+
width: 38px;
|
169 |
+
height: 38px;
|
170 |
+
border-radius: 50%;
|
171 |
+
margin-right: 10px;
|
172 |
+
}
|
173 |
+
.chat-message {
|
174 |
+
display: flex;
|
175 |
+
align-items: flex-start;
|
176 |
+
margin-bottom: 1rem;
|
177 |
}
|
178 |
+
.message-bubble {
|
179 |
+
background-color: #1f2937;
|
180 |
+
padding: 12px 16px;
|
181 |
border-radius: 12px;
|
182 |
+
max-width: 90%;
|
|
|
|
|
|
|
183 |
}
|
184 |
+
.chat-input {
|
185 |
+
background-color: #1f2937;
|
186 |
+
border: 1px solid #374151;
|
187 |
+
border-radius: 8px;
|
188 |
+
color: #e5e7eb;
|
189 |
+
padding: 0.75rem 1rem;
|
190 |
+
}
|
191 |
+
.gr-button.primary {
|
192 |
+
background: #2563eb;
|
193 |
color: white;
|
|
|
|
|
|
|
194 |
border-radius: 8px;
|
195 |
+
font-weight: 600;
|
196 |
}
|
197 |
+
.gr-button.primary:hover {
|
198 |
+
background: #1e40af;
|
199 |
}
|
200 |
""") as demo:
|
201 |
+
gr.Markdown("""<h2 style='color:#60a5fa'>π©Ί Patient History AI Assistant</h2><p>Upload a clinical Excel file and receive a structured diagnostic summary.</p>""")
|
202 |
+
with gr.Row():
|
203 |
+
with gr.Column(scale=3):
|
204 |
+
chatbot = gr.Chatbot(
|
205 |
+
label="Clinical Assistant",
|
206 |
+
height=700,
|
207 |
+
type="messages",
|
208 |
+
avatar_images=[
|
209 |
+
"https://ui-avatars.com/api/?name=AI&background=2563eb&color=fff&size=128",
|
210 |
+
"https://ui-avatars.com/api/?name=You&background=374151&color=fff&size=128"
|
211 |
+
]
|
212 |
+
)
|
213 |
+
with gr.Column(scale=1):
|
214 |
+
with gr.Row():
|
215 |
+
file_upload = gr.File(label="", file_types=[".xlsx"], elem_id="upload-btn")
|
216 |
+
analyze_btn = gr.Button("π§ Analyze", variant="primary")
|
217 |
+
report_output = gr.File(label="Download Report", visible=False, interactive=False)
|
218 |
+
|
219 |
+
chatbot_state = gr.State(value=[])
|
220 |
+
|
221 |
+
def update_ui(file, current_state):
|
222 |
+
messages, report_path = process_final_report(agent, file, current_state)
|
223 |
+
return messages, gr.update(visible=report_path is not None, value=report_path), messages
|
224 |
+
|
225 |
+
analyze_btn.click(fn=update_ui, inputs=[file_upload, chatbot_state], outputs=[chatbot, report_output, chatbot_state])
|
226 |
|
227 |
return demo
|
228 |
|
|
|
230 |
try:
|
231 |
agent = init_agent()
|
232 |
demo = create_ui(agent)
|
233 |
+
demo.launch(server_name="0.0.0.0", server_port=7860, allowed_paths=["/data/hf_cache/reports"], share=False)
|
234 |
except Exception as e:
|
235 |
print(f"Error: {str(e)}")
|
236 |
sys.exit(1)
|