Spaces:
Sleeping
Sleeping
Update backend.py
Browse files- backend.py +27 -52
backend.py
CHANGED
|
@@ -24,7 +24,8 @@ from bert import analyze_with_clinicalBert, classify_disease_and_severity, extra
|
|
| 24 |
from disease_links import diseases as disease_links
|
| 25 |
from disease_steps import disease_next_steps
|
| 26 |
from disease_support import disease_doctor_specialty, disease_home_care
|
| 27 |
-
|
|
|
|
| 28 |
|
| 29 |
model = genai.GenerativeModel('gemini-1.5-flash')
|
| 30 |
df = pd.read_csv("measurement.csv")
|
|
@@ -42,21 +43,7 @@ api = APIRouter(prefix="/api")
|
|
| 42 |
app.include_router(api)
|
| 43 |
|
| 44 |
|
| 45 |
-
'''app.add_middleware(
|
| 46 |
-
CORSMiddleware,
|
| 47 |
-
allow_origins=[
|
| 48 |
-
"http://localhost:8002"
|
| 49 |
-
"http://localhost:9000"
|
| 50 |
-
"http://localhost:5501"
|
| 51 |
-
],
|
| 52 |
-
allow_credentials=True,
|
| 53 |
-
allow_methods=["*"],
|
| 54 |
-
allow_headers=["*"],
|
| 55 |
-
)'''
|
| 56 |
-
|
| 57 |
-
|
| 58 |
app.mount("/app", StaticFiles(directory="web", html=True), name="web")
|
| 59 |
-
app.include_router(reports_router)
|
| 60 |
|
| 61 |
app.add_middleware(
|
| 62 |
CORSMiddleware,
|
|
@@ -95,13 +82,15 @@ try:
|
|
| 95 |
except Exception as e:
|
| 96 |
raise RuntimeError(f"Failed to configure Firebase: {e}")
|
| 97 |
|
| 98 |
-
class ChatRequest(BaseModel):
|
| 99 |
-
user_id: Optional[str] = "anonymous"
|
| 100 |
-
question: str
|
| 101 |
|
| 102 |
class ChatResponse(BaseModel):
|
| 103 |
answer: str
|
| 104 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 105 |
class ReportData(BaseModel):
|
| 106 |
user_id: str
|
| 107 |
reportDate: Optional[str] = None
|
|
@@ -169,7 +158,16 @@ def ocr_text_from_image(image_bytes: bytes) -> str:
|
|
| 169 |
print(response_text)
|
| 170 |
|
| 171 |
return response_text
|
| 172 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 173 |
try:
|
| 174 |
reports_ref = db.collection('users').document(request.user_id).collection('reports')
|
| 175 |
docs = reports_ref.order_by('timestamp', direction=firestore.Query.DESCENDING).limit(10).stream()
|
|
@@ -180,35 +178,8 @@ def get_past_reports_from_firestore(user_id: str):
|
|
| 180 |
history_text += f"Report from {report_data.get('timestamp', 'N/A')}:\n{report_data.get('ocr_text', 'No OCR text found')}\n\n"
|
| 181 |
except Exception as e:
|
| 182 |
history_text = "No past reports found for this user."
|
| 183 |
-
return history_text
|
| 184 |
-
|
| 185 |
-
def get_past_reports_from_sqllite(user_id: str):
|
| 186 |
-
try:
|
| 187 |
-
reports = db_fetch_reports(user_id=user_id, limit=10, offset=0)
|
| 188 |
-
|
| 189 |
-
history_text = ""
|
| 190 |
-
for report in reports:
|
| 191 |
-
history_text += f"Report from {report.get('report_date', 'N/A')}:\n{report.get('ocr_text', 'No OCR text found')}\n\n"
|
| 192 |
-
except Exception as e:
|
| 193 |
-
history_text = "No past reports found for this user."
|
| 194 |
-
return history_text
|
| 195 |
-
|
| 196 |
-
@app.post("/chat/", response_model=ChatResponse)
|
| 197 |
-
async def chat_endpoint(request: ChatRequest):
|
| 198 |
-
"""
|
| 199 |
-
Chatbot endpoint that answers questions based on the last analyzed document and user history.
|
| 200 |
-
"""
|
| 201 |
-
print("Received chat request for user:", request.user_id)
|
| 202 |
-
#history_text = get_past_reports_from_firestore(request.user_id)
|
| 203 |
-
history_text = get_past_reports_from_sqllite(request.user_id)
|
| 204 |
-
|
| 205 |
-
full_document_text = EXTRACTED_TEXT_CACHE + "\n\n" + "PAST REPORTS:\n" + history_text
|
| 206 |
-
|
| 207 |
-
if not full_document_text:
|
| 208 |
-
raise HTTPException(status_code=400, detail="No past reports or current data exists for this user")
|
| 209 |
-
|
| 210 |
|
| 211 |
-
|
| 212 |
|
| 213 |
try:
|
| 214 |
full_prompt = system_prompt_chat.format(
|
|
@@ -234,7 +205,6 @@ async def analyze(
|
|
| 234 |
filename = file.filename.lower()
|
| 235 |
detected_diseases = set()
|
| 236 |
ocr_full = ""
|
| 237 |
-
print("Received request for file:", filename)
|
| 238 |
if filename.endswith(".pdf"):
|
| 239 |
pdf_bytes = await file.read()
|
| 240 |
image_bytes_list = extract_images_from_pdf_bytes(pdf_bytes)
|
|
@@ -255,19 +225,22 @@ async def analyze(
|
|
| 255 |
return {"message": "Gemini model not available; please use BERT model."}
|
| 256 |
|
| 257 |
found_diseases = extract_non_negated_keywords(ocr_full)
|
|
|
|
| 258 |
past = detect_past_diseases(ocr_full)
|
|
|
|
| 259 |
|
| 260 |
for disease in found_diseases:
|
| 261 |
if disease in past:
|
| 262 |
severity = classify_disease_and_severity(disease)
|
| 263 |
detected_diseases.add(((f"{disease}(detected as historical condition, but still under risk.)"), severity))
|
|
|
|
| 264 |
else:
|
| 265 |
severity = classify_disease_and_severity(disease)
|
| 266 |
detected_diseases.add((disease, severity))
|
| 267 |
-
|
| 268 |
|
| 269 |
-
|
| 270 |
-
print("Detected diseases:",
|
| 271 |
ranges = analyze_measurements(ocr_full, df)
|
| 272 |
|
| 273 |
|
|
@@ -301,6 +274,7 @@ async def analyze(
|
|
| 301 |
next_steps_range = disease_next_steps.get(condition.lower(), ['Consult a doctor'])
|
| 302 |
specialist_range = disease_doctor_specialty.get(condition.lower(), "General Practitioner")
|
| 303 |
home_care_range = disease_home_care.get(condition.lower(), [])
|
|
|
|
| 304 |
|
| 305 |
condition_version = condition.upper()
|
| 306 |
severity_version = severity.upper()
|
|
@@ -314,11 +288,12 @@ async def analyze(
|
|
| 314 |
"info_link": link_range
|
| 315 |
})
|
| 316 |
|
| 317 |
-
|
| 318 |
ranges = analyze_measurements(ocr_full, df)
|
| 319 |
print(analyze_measurements(ocr_full, df))
|
| 320 |
# print ("Ranges is being printed", ranges)
|
| 321 |
historical_med_data = detect_past_diseases(ocr_full)
|
|
|
|
| 322 |
|
| 323 |
return {
|
| 324 |
"ocr_text": ocr_full.strip(),
|
|
|
|
| 24 |
from disease_links import diseases as disease_links
|
| 25 |
from disease_steps import disease_next_steps
|
| 26 |
from disease_support import disease_doctor_specialty, disease_home_care
|
| 27 |
+
import datetime
|
| 28 |
+
from typing import Optional, List
|
| 29 |
|
| 30 |
model = genai.GenerativeModel('gemini-1.5-flash')
|
| 31 |
df = pd.read_csv("measurement.csv")
|
|
|
|
| 43 |
app.include_router(api)
|
| 44 |
|
| 45 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 46 |
app.mount("/app", StaticFiles(directory="web", html=True), name="web")
|
|
|
|
| 47 |
|
| 48 |
app.add_middleware(
|
| 49 |
CORSMiddleware,
|
|
|
|
| 82 |
except Exception as e:
|
| 83 |
raise RuntimeError(f"Failed to configure Firebase: {e}")
|
| 84 |
|
|
|
|
|
|
|
|
|
|
| 85 |
|
| 86 |
class ChatResponse(BaseModel):
|
| 87 |
answer: str
|
| 88 |
|
| 89 |
+
|
| 90 |
+
class ChatRequest(BaseModel):
|
| 91 |
+
question: str
|
| 92 |
+
user_id: Optional[str] = "anonymous"
|
| 93 |
+
|
| 94 |
class ReportData(BaseModel):
|
| 95 |
user_id: str
|
| 96 |
reportDate: Optional[str] = None
|
|
|
|
| 158 |
print(response_text)
|
| 159 |
|
| 160 |
return response_text
|
| 161 |
+
|
| 162 |
+
@app.post("/chat/", response_model=ChatResponse)
|
| 163 |
+
async def chat_endpoint(request: ChatRequest):
|
| 164 |
+
"""
|
| 165 |
+
Chatbot endpoint that answers questions based on the last analyzed document and user history.
|
| 166 |
+
"""
|
| 167 |
+
global EXTRACTED_TEXT_CACHE
|
| 168 |
+
if not EXTRACTED_TEXT_CACHE:
|
| 169 |
+
raise HTTPException(status_code=400, detail="Please provide a document context by analyzing text first.")
|
| 170 |
+
|
| 171 |
try:
|
| 172 |
reports_ref = db.collection('users').document(request.user_id).collection('reports')
|
| 173 |
docs = reports_ref.order_by('timestamp', direction=firestore.Query.DESCENDING).limit(10).stream()
|
|
|
|
| 178 |
history_text += f"Report from {report_data.get('timestamp', 'N/A')}:\n{report_data.get('ocr_text', 'No OCR text found')}\n\n"
|
| 179 |
except Exception as e:
|
| 180 |
history_text = "No past reports found for this user."
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 181 |
|
| 182 |
+
full_document_text = EXTRACTED_TEXT_CACHE + "\n\n" + "PAST REPORTS:\n" + history_text
|
| 183 |
|
| 184 |
try:
|
| 185 |
full_prompt = system_prompt_chat.format(
|
|
|
|
| 205 |
filename = file.filename.lower()
|
| 206 |
detected_diseases = set()
|
| 207 |
ocr_full = ""
|
|
|
|
| 208 |
if filename.endswith(".pdf"):
|
| 209 |
pdf_bytes = await file.read()
|
| 210 |
image_bytes_list = extract_images_from_pdf_bytes(pdf_bytes)
|
|
|
|
| 225 |
return {"message": "Gemini model not available; please use BERT model."}
|
| 226 |
|
| 227 |
found_diseases = extract_non_negated_keywords(ocr_full)
|
| 228 |
+
print(f"CALLING FOUND DISEASES: {found_diseases}")
|
| 229 |
past = detect_past_diseases(ocr_full)
|
| 230 |
+
print(f"CALLING PAST DISEASES: {past}")
|
| 231 |
|
| 232 |
for disease in found_diseases:
|
| 233 |
if disease in past:
|
| 234 |
severity = classify_disease_and_severity(disease)
|
| 235 |
detected_diseases.add(((f"{disease}(detected as historical condition, but still under risk.)"), severity))
|
| 236 |
+
print(f"DETECTED DISEASES(PAST): {detected_diseases}")
|
| 237 |
else:
|
| 238 |
severity = classify_disease_and_severity(disease)
|
| 239 |
detected_diseases.add((disease, severity))
|
| 240 |
+
print(f"DETECTED DISEASES: {detected_diseases}")
|
| 241 |
|
| 242 |
+
print("OCR TEXT:", ocr_text)
|
| 243 |
+
print("Detected diseases:", found_diseases)
|
| 244 |
ranges = analyze_measurements(ocr_full, df)
|
| 245 |
|
| 246 |
|
|
|
|
| 274 |
next_steps_range = disease_next_steps.get(condition.lower(), ['Consult a doctor'])
|
| 275 |
specialist_range = disease_doctor_specialty.get(condition.lower(), "General Practitioner")
|
| 276 |
home_care_range = disease_home_care.get(condition.lower(), [])
|
| 277 |
+
print(f"HELLO!: {measurement}")
|
| 278 |
|
| 279 |
condition_version = condition.upper()
|
| 280 |
severity_version = severity.upper()
|
|
|
|
| 288 |
"info_link": link_range
|
| 289 |
})
|
| 290 |
|
| 291 |
+
print(ocr_full)
|
| 292 |
ranges = analyze_measurements(ocr_full, df)
|
| 293 |
print(analyze_measurements(ocr_full, df))
|
| 294 |
# print ("Ranges is being printed", ranges)
|
| 295 |
historical_med_data = detect_past_diseases(ocr_full)
|
| 296 |
+
print("***End of Code***")
|
| 297 |
|
| 298 |
return {
|
| 299 |
"ocr_text": ocr_full.strip(),
|