clip-protocol 2.2.14__tar.gz → 2.3__tar.gz

{clip_protocol-2.2.14 → clip_protocol-2.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: clip_protocol
-Version: 2.2.14
+Version: 2.3
 Summary: Protocol to ensure the privatization of 
 Project-URL: Homepage, https://github.com/martaajonees/Local_Privacy
 Project-URL: Issues, https://github.com/martaajonees/Local_Privacy/issues

clip_protocol-2.3/evaluation/AOI datasets/process.py

@@ -0,0 +1,57 @@
+import pandas as pd
+import glob
+import os
+
+# Carpeta actual donde está el script
+current_folder = os.getcwd()
+
+# Buscar todos los archivos .xlsx en la carpeta actual
+xlsx_files = glob.glob(os.path.join(current_folder, "*.xlsx"))
+
+for file_path in xlsx_files:
+    try:
+        # Leer el archivo
+        df_temp = pd.read_excel(file_path)
+
+        if any(col.startswith("Unnamed") for col in df_temp.columns):
+            df = pd.read_excel(file_path, header=1)  
+        else:
+            df = df_temp
+
+        # --- MEJORA: Limpiar nombres de columnas ---
+        # Eliminamos espacios en blanco extra y pasamos a minúsculas para comparar fácilmente
+        df.columns = [str(c).strip() for c in df.columns]
+        cols_lower = [c.lower() for c in df.columns]
+
+        # Buscamos los índices de las columnas que necesitamos (sin importar mayúsculas)
+        target_p = 'participant'
+        target_a = 'aoi hit'
+
+        if target_p in cols_lower and target_a in cols_lower:
+            # Identificar los nombres reales de las columnas en este archivo específico
+            real_col_p = df.columns[cols_lower.index(target_p)]
+            real_col_a = df.columns[cols_lower.index(target_a)]
+
+            # Obtener el nombre del archivo sin extensión para el participante
+            participant_name = os.path.splitext(os.path.basename(file_path))[0]
+
+            # Reemplazar valores y filtrar
+            df[real_col_p] = participant_name
+            
+            # Mantener solo las dos columnas y limpiar nulos
+            df = df[[real_col_p, real_col_a]]
+            df = df[df[real_col_a].notna()]
+
+            # Renombrar a los nombres finales deseados
+            df.columns = ['Participant', 'AOI hit']
+
+            # Sobrescribir el archivo original
+            df.to_excel(file_path, index=False)
+            print(f"✅ Procesado: {os.path.basename(file_path)}")
+        else:
+            print(f"❌ Columnas no encontradas en {os.path.basename(file_path)}. Columnas detectadas: {list(df.columns)}")
+
+    except Exception as e:
+        print(f"⚠️ Error procesando {os.path.basename(file_path)}: {e}")
+
+print("\nProceso finalizado.")

{clip_protocol-2.2.14 → clip_protocol-2.3}/evaluation/experiment_1.py

@@ -24,38 +24,6 @@ def run_command(e, k, m, df, privacy_method):
 
     return compute_error_table(get_real_frequency(df), df_estimated, 2), df_estimated
 
-
-def plot_latex(errors, path):
-    lines = [
-        r"\begin{figure}[h]",
-        r"\centering",
-        r"\begin{tikzpicture}",
-        r"\begin{axis}[",
-        r"  xlabel={$\epsilon$}, ylabel={Error},",
-        r"  legend style={at={(0.5,-0.15)}, anchor=north,legend columns=-1},",
-        r"  xmin=0, grid=major, width=12cm, height=8cm,",
-        r"  cycle list name=color list,",
-        r"]"
-    ]
-
-    for metric, points in errors.items():
-        name = "Lp Norm" if metric == "Lρ Norm" else metric
-        lines.append(r"\addplot coordinates {")
-        lines += [f"  ({eps}, {err})" for eps, err in sorted(points)]
-        lines.append(r"};")
-        lines.append(fr"\addlegendentry{{{name}}}")
-
-    lines += [
-        r"\end{axis}",
-        r"\end{tikzpicture}",
-        r"\caption{Evolución del error por métrica en función del parámetro $\epsilon$}",
-        r"\end{figure}"
-    ]
-
-    with open(path, "w") as f:
-        f.write("\n".join(lines))
-    print(f"✅ LaTeX graph saved to {path}")
-
 def run_experiment1(df, privacy_method):
     k = int(input("🔑 Enter k value: "))
     m = int(input("🔢 Enter m value: "))
@@ -65,13 +33,13 @@ def run_experiment1(df, privacy_method):
     epsilons = [round(e, 1) for e in list(reversed([x * 0.5 for x in range(1, 21)])) + [0.4, 0.3, 0.2, 0.1]]
 
     for eps in epsilons:
-        table, _ = run_command(eps, k, m, df, method)
+        print(f"Running with epsilon: {eps}")
+        table, _ = run_command(eps, k, m, df, privacy_method)
         for metric, val in table:
             error_history.setdefault(metric, []).append((eps, val))
     
     df = pd.DataFrame(error_history)
-    df.to_csv(f"figures/table_experiment_1_{privacy_method}.csv", index=False)
-    plot_latex(error_history, f"figures/experiment_1_{privacy_method}.tex")
+    df.to_csv(f"table_experiment_1_{privacy_method}.csv", index=False)
         
 if __name__ == "__main__":
     parser = argparse.ArgumentParser(description="Run experiment 1")
@@ -79,8 +47,8 @@ if __name__ == "__main__":
     args = parser.parse_args()
     
     distribution = input(" Enter the distribution 1/2/3/4: ")
-
-    pattern = f"aoi-hits-d{distribution}-5000"
+    
+    pattern = f"SynLog-5000-d{distribution}"
     matching_files = [f for f in os.listdir(args.f) if pattern in f and f.endswith(".xlsx")]
 
     file_path = os.path.join(args.f, matching_files[0])

{clip_protocol-2.2.14 → clip_protocol-2.3}/evaluation/experiment_2.py

@@ -115,7 +115,7 @@ def run_experiment_2(datasets_by_size, params):
                   "Iteraciones PCMeS", "Tiempo de ejecución PCMeS"]
     df_pivot = df_pivot[final_cols]
 
-    df_pivot.to_csv("figures/table_experiment_2.csv", index=False)
+    df_pivot.to_csv("table_experiment_2.csv", index=False)
 
 
 if __name__ == "__main__":
@@ -133,7 +133,8 @@ if __name__ == "__main__":
 
     datasets = {}
     for size in sizes:
-        pattern = f"aoi-hits-d{distribution}-{size}"
+        
+        pattern = f"SynLog-{size}-d{distribution}"
         file_path = os.path.join(args.f, pattern + ".xlsx")
         header = 1 if "Unnamed" in pd.read_excel(file_path, nrows=1).columns[0] else 0
         df = pd.read_excel(file_path, header=header)

{clip_protocol-2.2.14 → clip_protocol-2.3}/evaluation/experiment_3.py

@@ -66,7 +66,8 @@ def optimize_e(k, m, df, e_r, privacy_level, error_value, tolerance, privacy_met
 def run_experiment_3(datasets, params):
     error_value = 0.05
     tolerance = 0.01
-    privacy_level = "high"
+    # privacy_level = "high"
+    privacy_level = "low"
 
     for method in ["PCMeS", "PHCMS"]:
         row_apple = {"Método": "Método de Apple"}
@@ -89,7 +90,7 @@ def run_experiment_3(datasets, params):
             row_clip[size] = f"{epsilon:.2f} / {pe_error:.2f}"
         
         df_result = pd.DataFrame([row_apple, row_clip])
-        df_result.to_csv(f"figures/table_experiment_3_{method}.csv", index=False)
+        df_result.to_csv(f"table_experiment_3_{method}.csv", index=False)
         
 
 if __name__ == "__main__":
@@ -107,11 +108,11 @@ if __name__ == "__main__":
 
     datasets = {}
     for size in sizes:
-        pattern = f"aoi-hits-d{distribution}-{size}"
+        pattern = f"SynLog-{size}-d{distribution}"
         file_path = os.path.join(args.f, pattern + ".xlsx")
         header = 1 if "Unnamed" in pd.read_excel(file_path, nrows=1).columns[0] else 0
         df = pd.read_excel(file_path, header=header)
         datasets[size] = df
         
     
-    run_experiment_3(datasets, params)
+    run_experiment_3(datasets, params)

{clip_protocol-2.2.14 → clip_protocol-2.3}/evaluation/experiment_4.py

@@ -84,7 +84,7 @@ def run_experiment_4(datasets, params):
             cleaned_table = [[col[0], col[1].replace('%', '') if isinstance(col[1], str) else col[1]] for col in filtered_table]
 
             error_by_aoi = pd.DataFrame(cleaned_table, columns=['AOI', 'Error'])
-            path_individual = f"figures/experimet_4_d{distribution}_{method}.csv"
+            path_individual = f"experimet_4_d{distribution}_{method}.csv"
             error_by_aoi.to_csv(path_individual, index=False)
         
 if __name__ == "__main__":
@@ -101,7 +101,7 @@ if __name__ == "__main__":
 
     datasets = {}
     for distribution in distributions:
-        pattern = f"aoi-hits-d{distribution}-5000"
+        pattern = f"SynLog-5000-d{distribution}"
         file_path = os.path.join(args.f, pattern + ".xlsx")
         header = 1 if "Unnamed" in pd.read_excel(file_path, nrows=1).columns[0] else 0
         df = pd.read_excel(file_path, header=header)

{clip_protocol-2.2.14 → clip_protocol-2.3}/evaluation/experiment_5.py

@@ -149,10 +149,10 @@ def run_experiment_5(datasets, privatized_path):
             privatized_data.to_csv(filename, index=False)
     
     performance_df = pd.DataFrame(performance_records)
-    performance_df.to_csv("figures/experiment_5.csv", index=False)
+    performance_df.to_csv("experiment_5.csv", index=False)
 
     size_df = pd.DataFrame(size_comparison_records)
-    size_df.to_csv("figures/experiment_5_size_comparison.csv", index=False)
+    size_df.to_csv("experiment_5_size_comparison.csv", index=False)
 
 def load_excel_with_header_check(filepath):
     try:

{clip_protocol-2.2.14 → clip_protocol-2.3}/evaluation/figures/experiment_2_params.json

@@ -6,7 +6,7 @@
   },
   "PHCMS": {
     "k": 145,
-    "m": 128,
+    "m": 256,
     "e_r": 8.0
   }
-}
+}

{clip_protocol-2.2.14 → clip_protocol-2.3}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "clip_protocol"
-version = "2.2.14"
+version = "2.3"
 authors = [
   { name="Marta Jones González", email="martajon10@gmail.com" },
 ]

{clip_protocol-2.2.14 → clip_protocol-2.3}/src/clip_protocol/main/setup.py

@@ -76,6 +76,7 @@ class Setup:
         self.df['value'] = self.df['value'].astype(str).apply(lambda x: x.strip())
         self.df = self.df[self.df['value'] != '-']
         self.df = self.df[self.df['value'].str.contains(r'\w', na=False)]
+        self.df = self.df.sample(frac=1, random_state=None).reset_index(drop=True)
         self.N = len(self.df)
     
     def run_command(self, e, k, m):

clip_protocol-2.2.14/evaluation/datasets-article/process.py

@@ -1,46 +0,0 @@
-import os
-import glob
-import re
-import pandas as pd
-import random
-import string
-import hashlib
-
-
-# ID único que quieres usar
-UNIQUE_USER_ID= ''.join(random.choices(string.ascii_uppercase + string.digits, k=5))
-
-
-# Buscar todos los archivos xlsx en el directorio actual
-files = glob.glob("*.xlsx")
-
-for file in files:
-    print(f"Procesando {file}...")
-    
-    # Leer el Excel
-    df = pd.read_excel(file)
-    
-    # 1️⃣ Cambiar todos los user_id a uno único
-    if "user_id" in df.columns:
-        df["user_id"] = UNIQUE_USER_ID
-    else:
-        print(f"⚠️  No se encontró columna 'user_id' en {file}")
-    
-    # 2️⃣ Renombrar la segunda columna a 'events'
-    if len(df.columns) >= 2:
-        cols = list(df.columns)
-        cols[1] = "events"
-        df.columns = cols
-    else:
-        print(f"⚠️  El archivo {file} no tiene al menos 2 columnas")
-    
-    # 3️⃣ Reemplazar subevent_X → eX dentro de la columna events
-    if "events" in df.columns:
-        df["events"] = df["events"].astype(str).apply(
-            lambda x: re.sub(r"subevent_(\d+)", r"e\1", x)
-        )
-    
-    # Guardar sobrescribiendo el archivo
-    df.to_excel(file, index=False)
-
-print("✅ Procesamiento terminado.")