Update app.py

app.py

@@ -8,6 +8,51 @@ from Gradio_UI import GradioUI
 
 # Custom Tool to fetch datasets related to body parts or imaging types
 
+# @tool
+# def my_custom_tool(arg1: str, arg2: int) -> str:
+#     """
+#     Search and retrieve publicly available medical datasets from Hugging Face based on any medical-related keyword.
+
+#     Args:
+#         arg1: A keyword related to medical data (e.g., 'cancer', 'diabetes', 'CT scan', 'radiology', 'dermoscopy').
+#         arg2: The maximum number of datasets to retrieve.
+
+#     Returns:
+#         A list of dataset names matching the search query, or a message stating that no datasets were found
+#         or that the keyword is not medically relevant.
+#     """
+#     try:
+#         keyword = arg1.strip().lower()
+#         limit = int(arg2)
+
+#         # ✅ Define a basic list of medically relevant terms
+#         medical_terms = [
+#             "skin", "brain", "cancer", "breast cancer", "prostate cancer", "stomach", "tumor", "mri", "ct", "xray", "ultrasound",
+#             "diabetes", "pneumonia", "covid", "lesion", "radiology", "pathology",
+#             "lung", "chest", "abdomen", "spine", "bone", "stroke", "eczema", "melanoma",
+#             "eye", "retina", "dermoscopy", "cardiology", "infection", "biopsy", "tooth",
+#             "toothache", "dental", "ear", "wrist", "hand", "leg", "arm", "heart"
+#         ]
+
+#         # ✅ Check if the keyword looks medically relevant
+#         if not any(term in keyword for term in medical_terms):
+#             return f"'{arg1}' does not appear to be a medical term."
+
+#         # ✅ Proceed to fetch datasets
+#         response = requests.get(
+#             f"https://huggingface.co/api/datasets?search={keyword}&limit={limit}"
+#         )
+#         response.raise_for_status()
+#         datasets = response.json()
+
+#         if not datasets:
+#             return f"No medical datasets found for '{arg1}'."
+
+#         results = [f"- {ds.get('id', 'Unknown')}" for ds in datasets[:limit]]
+#         return f"Medical datasets related to '{arg1}':\n" + "\n".join(results)
+
+#     except Exception as e:
+#         return f"Error searching medical datasets for '{arg1}': {str(e)}"
 @tool
 def my_custom_tool(arg1: str, arg2: int) -> str:
     """
@@ -18,36 +63,58 @@ def my_custom_tool(arg1: str, arg2: int) -> str:
         arg2: The maximum number of datasets to retrieve.
 
     Returns:
-        A list of dataset names matching the search query, or a message stating that no datasets were found
-        or that the keyword is not medically relevant.
+        A list of dataset names matching the search query, or a message stating that no datasets were found.
     """
     try:
         keyword = arg1.strip().lower()
         limit = int(arg2)
 
-        # ✅ Define a basic list of medically relevant terms
+        # Define a basic list of medically relevant terms
         medical_terms = [
-            "skin", "brain", "cancer", "breast cancer", "prostate cancer", "stomach", "tumor", "mri", "ct", "xray", "ultrasound",
-            "diabetes", "pneumonia", "covid", "lesion", "radiology", "pathology",
-            "lung", "chest", "abdomen", "spine", "bone", "stroke", "eczema", "melanoma",
-            "eye", "retina", "dermoscopy", "cardiology", "infection", "biopsy", "tooth",
-            "toothache", "dental", "ear", "wrist", "hand", "leg", "arm", "heart"
+            # Anatomy / Body Parts
+            "skin", "brain", "lung", "chest", "abdomen", "spine", "bone", "heart", "liver", "kidney",
+            "bladder", "stomach", "colon", "rectum", "esophagus", "pancreas", "breast", "ear", "eye", 
+            "retina", "tooth", "tongue", "jaw", "neck", "wrist", "hand", "leg", "arm", "shoulder", "pelvis",
+        
+            # Diseases / Conditions
+            "cancer", "tumor", "stroke", "diabetes", "pneumonia", "covid", "asthma", "eczema", "melanoma",
+            "hypertension", "alzheimer", "parkinson", "arthritis", "scoliosis", "epilepsy", "glaucoma",
+            "ulcer", "hepatitis", "leukemia", "lymphoma", "tuberculosis", "anemia", "obesity", "depression",
+        
+            # Imaging Modalities
+            "mri", "ct", "xray", "x-ray", "ultrasound", "pet", "fmri", "mammo", "angiography", "radiography",
+            "echocardiogram", "spect", "dermoscopy", "colonoscopy", "endoscopy", "biopsy", "histopathology",
+        
+            # Medical Specialties
+            "radiology", "pathology", "oncology", "cardiology", "neurology", "dermatology", "dentistry",
+            "ophthalmology", "urology", "orthopedics", "gastroenterology", "pulmonology", "nephrology",
+        
+            # Symptoms / Signs
+            "lesion", "infection", "fever", "pain", "inflammation", "rash", "headache", "swelling", 
+            "cough", "seizure", "dizziness", "vomiting", "diarrhea", "nausea", "fatigue", "itching",
+        
+            # Common Specific Diseases
+            "breast cancer", "prostate cancer", "lung cancer", "skin cancer", "colon cancer", 
+            "brain tumor", "liver cancer", "cervical cancer", "bladder cancer", "thyroid cancer"
         ]
 
-        # ✅ Check if the keyword looks medically relevant
+
+        # Check if keyword is in known medical terms
         if not any(term in keyword for term in medical_terms):
-            return f"'{arg1}' does not appear to be a medical term."
+            return f"No medical datasets found for '{arg1}'."
 
-        # ✅ Proceed to fetch datasets
+        # Fetch datasets from Hugging Face
         response = requests.get(
             f"https://huggingface.co/api/datasets?search={keyword}&limit={limit}"
         )
         response.raise_for_status()
         datasets = response.json()
 
+        # Return message if no datasets found
         if not datasets:
             return f"No medical datasets found for '{arg1}'."
 
+        # Collect and return dataset names
         results = [f"- {ds.get('id', 'Unknown')}" for ds in datasets[:limit]]
         return f"Medical datasets related to '{arg1}':\n" + "\n".join(results)