Update app.py

app.py

@@ -1,5 +1,4 @@
-from fastapi import FastAPI
-from pydantic import BaseModel
+from fastapi import FastAPI, UploadFile, File
 import easyocr
 import cv2
 import numpy as np
@@ -14,36 +13,18 @@ app = FastAPI()
 
 reader = easyocr.Reader(['en'])
 
-# =========================
-# REQUEST MODEL
-# =========================
-
-class ImageRequest(BaseModel):
-    image_path: str
-
 # =========================
 # IMAGE QUALITY CHECKS
 # =========================
 
 def is_blurry(image):
-
-    gray = cv2.cvtColor(
-        image,
-        cv2.COLOR_BGR2GRAY
-    )
-
-    variance = cv2.Laplacian(
-        gray,
-        cv2.CV_64F
-    ).var()
-
+    gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)
+    variance = cv2.Laplacian(gray, cv2.CV_64F).var()
     return variance < 100
 
 
 def is_dark(image):
-
     brightness = np.mean(image)
-
     return brightness < 50
 
 
@@ -52,13 +33,8 @@ def is_dark(image):
 # =========================
 
 def extract_text(image_path):
-
     results = reader.readtext(image_path)
-
-    text = " ".join(
-        [r[1] for r in results]
-    ).lower()
-
+    text = " ".join([r[1] for r in results]).lower()
     return text
 
 
@@ -72,11 +48,7 @@ def detect_document(text):
     text = text.lower().strip()
 
     # REMOVE EXTRA SYMBOLS
-    cleaned_text = re.sub(
-        r'[^a-zA-Z0-9\s-]',
-        ' ',
-        text
-    )
+    cleaned_text = re.sub(r'[^a-zA-Z0-9\s-]', ' ', text)
 
     # SPLIT WORDS
     words = cleaned_text.split()
@@ -90,11 +62,8 @@ def detect_document(text):
     ]
 
     for pattern in garbage_patterns:
-
         for word in words:
-
             if re.match(pattern, word):
-
                 if len(words) <= 2:
                     return {
                         "document_type": "unknown",
@@ -120,12 +89,10 @@ def detect_document(text):
     matched_keywords = []
 
     for keyword in nin_keywords:
-
         if keyword in cleaned_text:
             matched_keywords.append(keyword)
 
     if len(matched_keywords) > 0:
-
         return {
             "document_type": "nin",
             "confidence": 95,
@@ -145,12 +112,10 @@ def detect_document(text):
     matched_keywords = []
 
     for keyword in passport_keywords:
-
         if keyword in cleaned_text:
             matched_keywords.append(keyword)
 
     if len(matched_keywords) > 0:
-
         return {
             "document_type": "passport",
             "confidence": 94,
@@ -172,12 +137,10 @@ def detect_document(text):
     matched_keywords = []
 
     for keyword in license_keywords:
-
         if keyword in cleaned_text:
             matched_keywords.append(keyword)
 
     if len(matched_keywords) >= 2:
-
         return {
             "document_type": "drivers_license",
             "confidence": 92,
@@ -198,12 +161,10 @@ def detect_document(text):
     matched_keywords = []
 
     for keyword in voter_keywords:
-
         if keyword in cleaned_text:
             matched_keywords.append(keyword)
 
     if len(matched_keywords) > 0:
-
         return {
             "document_type": "voters_card",
             "confidence": 90,
@@ -211,7 +172,7 @@ def detect_document(text):
         }
 
     # =========================
-    # ELECTRICITY / UTILITY BILL
+    # ELECTRICITY COMPANIES
     # =========================
 
     electricity_keywords = [
@@ -246,7 +207,7 @@ def detect_document(text):
         "yedc",
         "yola electricity",
 
-        # Common terms
+        # Common Nigerian utility terms
         "prepaid",
         "postpaid",
         "disco",
@@ -257,12 +218,10 @@ def detect_document(text):
     matched_keywords = []
 
     for keyword in electricity_keywords:
-
         if keyword in cleaned_text:
             matched_keywords.append(keyword)
 
     if len(matched_keywords) > 0:
-
         return {
             "document_type": "utility_bill",
             "confidence": 90,
@@ -274,7 +233,6 @@ def detect_document(text):
     # =========================
 
     bank_keywords = [
-
         "account statement",
         "statement of account",
         "transaction",
@@ -301,12 +259,10 @@ def detect_document(text):
     matched_keywords = []
 
     for keyword in bank_keywords:
-
         if keyword in cleaned_text:
             matched_keywords.append(keyword)
 
     if len(matched_keywords) > 0:
-
         return {
             "document_type": "bank_statement",
             "confidence": 91,
@@ -318,7 +274,6 @@ def detect_document(text):
     # =========================
 
     tenancy_keywords = [
-
         "tenancy agreement",
         "landlord",
         "tenant",
@@ -331,12 +286,10 @@ def detect_document(text):
     matched_keywords = []
 
     for keyword in tenancy_keywords:
-
         if keyword in cleaned_text:
             matched_keywords.append(keyword)
 
     if len(matched_keywords) > 0:
-
         return {
             "document_type": "tenancy_agreement",
             "confidence": 89,
@@ -348,7 +301,6 @@ def detect_document(text):
     # =========================
 
     vehicle_keywords = [
-
         "toyota",
         "honda",
         "lexus",
@@ -371,7 +323,6 @@ def detect_document(text):
     matched_keywords = []
 
     for keyword in vehicle_keywords:
-
         if keyword in cleaned_text:
             matched_keywords.append(keyword)
 
@@ -380,7 +331,6 @@ def detect_document(text):
     # =========================
 
     nigeria_states = [
-
         "lagos",
         "abuja",
         "kano",
@@ -422,7 +372,6 @@ def detect_document(text):
     state_matches = []
 
     for state in nigeria_states:
-
         if state in cleaned_text:
             state_matches.append(state)
 
@@ -431,7 +380,6 @@ def detect_document(text):
     # =========================
 
     plate_patterns = [
-
         r"[A-Z]{3}-?\d{3}[A-Z]{2}",
         r"[A-Z]{2}\d{3}[A-Z]{3}",
         r"[A-Z]{3}\s\d{3}\s[A-Z]{2}"
@@ -440,16 +388,10 @@ def detect_document(text):
     detected_plate = None
 
     for pattern in plate_patterns:
-
-        plate_match = re.search(
-            pattern,
-            cleaned_text.upper()
-        )
+        plate_match = re.search(pattern, cleaned_text.upper())
 
         if plate_match:
-
             detected_plate = plate_match.group()
-
             break
 
     # =========================
@@ -457,18 +399,14 @@ def detect_document(text):
     # =========================
 
     if detected_plate:
-
         return {
             "document_type": "vehicle_plate",
             "confidence": 97,
-            "matched_keywords": [
-                detected_plate
-            ] + state_matches
+            "matched_keywords": [detected_plate] + state_matches
         }
 
     # VEHICLE WITHOUT CLEAR PLATE
     if len(matched_keywords) > 0:
-
         return {
             "document_type": "vehicle_image",
             "confidence": 75,
@@ -488,7 +426,6 @@ def detect_document(text):
 
 @app.get("/")
 def home():
-
     return {
         "success": True,
         "message": "Document Validation API Running",
@@ -510,28 +447,18 @@ def home():
 # =========================
 
 @app.post("/validate")
-async def validate_document(
-    request: ImageRequest
-):
+async def validate_document(file: UploadFile = File(...)):
 
     try:
 
-        image_path = request.image_path
-
         # =========================
-        # CHECK FILE EXISTS
+        # SAVE FILE
         # =========================
 
-        if not os.path.exists(image_path):
+        image_path = "temp.jpg"
 
-            return {
-                "success": False,
-                "message": "Image not found",
-                "reason": (
-                    "The provided image path "
-                    "does not exist."
-                )
-            }
+        with open(image_path, "wb") as f:
+            f.write(await file.read())
 
         # =========================
         # READ IMAGE
@@ -540,16 +467,15 @@ async def validate_document(
         image = cv2.imread(image_path)
 
         if image is None:
-
             return {
                 "success": False,
                 "message": "Invalid image",
                 "reason": (
-                    "The file could not be "
-                    "read as an image."
+                    "The uploaded file could not "
+                    "be read as an image."
                 ),
                 "suggestion": (
-                    "Provide a valid JPG or PNG image."
+                    "Upload a valid JPG or PNG image."
                 )
             }
 
@@ -558,14 +484,12 @@ async def validate_document(
         # =========================
 
         if is_blurry(image):
-
             return {
                 "success": False,
                 "message": "Image rejected",
-                "reason": "The image is blurry.",
+                "reason": "The uploaded image is blurry.",
                 "suggestion": (
-                    "Retake the photo with "
-                    "better focus."
+                    "Retake the photo with better focus."
                 )
             }
 
@@ -574,13 +498,12 @@ async def validate_document(
         # =========================
 
         if is_dark(image):
-
             return {
                 "success": False,
                 "message": "Image rejected",
-                "reason": "The image is too dark.",
+                "reason": "The uploaded image is too dark.",
                 "suggestion": (
-                    "Use better lighting."
+                    "Take the photo in a brighter environment."
                 )
             }
 
@@ -595,16 +518,16 @@ async def validate_document(
         # =========================
 
         if len(text.strip()) == 0:
-
             return {
                 "success": False,
                 "message": "Document rejected",
                 "reason": (
-                    "No readable text was detected."
+                    "No readable text was detected "
+                    "in the image."
                 ),
                 "suggestion": (
-                    "Ensure the document is "
-                    "clear and fully visible."
+                    "Ensure the document is clear "
+                    "and fully visible."
                 )
             }
 
@@ -619,13 +542,12 @@ async def validate_document(
         # =========================
 
         if document_result is None:
-
             return {
                 "success": False,
                 "message": "Document rejected",
                 "reason": (
-                    "The uploaded image does not "
-                    "match any supported document type."
+                    "The uploaded image does not match "
+                    "any supported document type."
                 ),
                 "supported_documents": [
                     "National ID (NIN)",
@@ -676,17 +598,11 @@ async def validate_document(
             "reason": str(e)
         }
 
+    finally:
 
-# =========================
-# RUN SERVER
-# =========================
-
-if __name__ == "__main__":
-
-    import uvicorn
+        # =========================
+        # CLEAN TEMP FILE
+        # =========================
 
-    uvicorn.run(
-        app,
-        host="0.0.0.0",
-        port=7860
-    )
+        if os.path.exists("temp.jpg"):
+            os.remove("temp.jpg")