refactor directories for better understanding of the project

Author: mtiessler

backend/dendogram_controller.py

@@ -91,7 +91,7 @@ def generate_dendogram_from_csv():
         return make_response("CSV file is required", 400)
 
     file = request.files['file']
-    if not file.filename.endswith('.csv'):
+    if not file.filename.endswith('.model_embeddings'):
         return make_response("File must be a CSV", 400)
 
     features = []

backend/dendogram_service.py

@@ -13,17 +13,17 @@
 
 load_dotenv()
 def preprocessed_app(app_name):
-    file_path = f"data/Stage 2 - Hierarchical Clustering/preprocessed_jsons/{app_name}Features.json"
+    file_path = f"data/Stage 2 - Hierarchical Clustering/preprocessed_features_jsons/{app_name}Features.json"
     return os.path.exists(file_path) and os.path.getsize(file_path) > 0
 
 def save_preprocessed_features(features, app_name):
-    file_path = f"data/Stage 2 - Hierarchical Clustering/preprocessed_jsons/{app_name}Features.json"
+    file_path = f"data/Stage 2 - Hierarchical Clustering/preprocessed_features_jsons/{app_name}Features.json"
     os.makedirs(os.path.dirname(file_path), exist_ok=True)
     with open(file_path, "w") as json_file:
         json.dump(features, json_file)
 
 def load_saved_preprocessed_features(app_name):
-    file_path = os.path.join(BASE_DIR, "data", "Stage 2 - Hierarchical Clustering", "preprocessed_jsons", f"{app_name}Features.json")
+    file_path = os.path.join(BASE_DIR, "data", "Stage 2 - Hierarchical Clustering", "preprocessed_features_jsons", f"{app_name}Features.json")
     if not os.path.exists(file_path):
         return None
     with open(file_path, "r") as json_file:

backend/preprocessing_service.py

@@ -33,17 +33,17 @@ def preprocess():
         return jsonify({"error": str(e)}), 500
 
 def preprocessed_app(app_name):
-    file_path = f"static/preprocessed_jsons/{app_name}Features.json"
+    file_path = f"data/Stage 3 - Topic Modelling/preprocessed_features_jsons/{app_name}Features.json"
     return os.path.exists(file_path) and os.path.getsize(file_path) > 0
 
 def save_preprocessed_features(features, app_name):
-    file_path = f"static/preprocessed_jsons/{app_name}Features.json"
+    file_path = f"data/Stage 3 - Topic Modelling/preprocessed_features_jsons/{app_name}Features.json"
     os.makedirs(os.path.dirname(file_path), exist_ok=True)
     with open(file_path, "w") as json_file:
         json.dump(features, json_file)
 
 def load_saved_preprocessed_features(app_name):
-    file_path = f"static/preprocessed_jsons/{app_name}Features.json"
+    file_path = f"data/Stage 3 - Topic Modelling/preprocessed_features_jsons/{app_name}Features.json"
     if os.path.exists(file_path):
         with open(file_path, "r") as json_file:
             return json.load(json_file)

backend/utils.py

@@ -11,8 +11,8 @@
 STAGE_2_MODEL_DIRECTORY_PATH = os.path.join(STAGE_2_OUTPUT_PATH)
 STAGE_3_MODEL_DIRECTORY_PATH = os.path.join(STAGE_3_INPUT_PATH)
 
-MODEL_DIRECTORY_CSV_PATH = os.path.join(STAGE_2_OUTPUT_PATH, 'csv')
-MODEL_DIRECTORY_CSV_EMBEDDINGS_PATH = os.path.join('static', 'csv', 'embeddings')
+MODEL_DIRECTORY_CSV_PATH = os.path.join(STAGE_2_OUTPUT_PATH, 'model_embeddings')
+MODEL_DIRECTORY_CSV_EMBEDDINGS_PATH = os.path.join('data', 'model_embeddings', 'embeddings')
 
 class Utils:
     @staticmethod

backend/visualization_service.py

@@ -49,7 +49,7 @@ def generate_dynamic_label(cluster_labels):
     unique_labels = list(set(cluster_labels))
     few_shot_input_text = (
         "Generate a single concise label summarizing the following actions:\n\n"
-        "Examples:\n"
+        "Input_Examples:\n"
         "Video meeting, online meeting, team video chat, conference call\n"
         "Label: Virtual Team Communication\n\n"
         "Secure chat, encrypted messaging, private message\n"

client/dendogram_generation.py cli-client/dendogram_generation.pyclient/dendogram_generation.py renamed to cli-client/dendogram_generation.py

@@ -11,7 +11,7 @@
 obj_weights = [0.9, 0.75, 0.669, 0.5, 0.25, 0.1]
 affinity_models = ['bert', 'tf-idf', 'paraphrase']
 
-json_file_path = "body.json"
+json_file_path = "small_whatsapp_example.json"
 with open(json_file_path, 'r') as json_file:
     json_data = json.load(json_file)
 

client/dynamic_visualizator.py cli-client/dynamic_visualizator.pyclient/dynamic_visualizator.py renamed to cli-client/dynamic_visualizator.py

@@ -60,7 +60,7 @@ def process_folder(input_folder, output_folder):
 
     # Process each CSV file in the input folder
     for file_name in os.listdir(input_folder):
-        if file_name.endswith('.csv'):
+        if file_name.endswith('.model_embeddings'):
             input_file = os.path.join(input_folder, file_name)
             output_file = os.path.join(output_folder, file_name)
             print(f"Processing {input_file} -> {output_file}")