aait 2.3.12.2__tar.gz → 2.3.12.3__tar.gz

aait-2.3.12.3/PKG-INFO

@@ -0,0 +1,32 @@
+Metadata-Version: 2.1
+Name: aait
+Version: 2.3.12.3
+Summary: Advanced Artificial Intelligence Tools is a package meant to develop and enable advanced AI functionalities in Orange
+Home-page: 
+Author: Orange community
+Author-email: 
+License: UNKNOWN
+Keywords: orange3 add-on
+Platform: UNKNOWN
+License-File: License.txt
+Requires-Dist: torch
+Requires-Dist: sentence-transformers==5.0.0
+Requires-Dist: gpt4all[all]==2.8.2
+Requires-Dist: sacremoses==0.1.1
+Requires-Dist: transformers==4.51.3
+Requires-Dist: sentencepiece==0.2.0
+Requires-Dist: optuna
+Requires-Dist: spacy==3.7.6
+Requires-Dist: markdown
+Requires-Dist: python-multipart
+Requires-Dist: PyMuPDF==1.24.14
+Requires-Dist: chonkie==0.4.1
+Requires-Dist: GPUtil==1.4.0
+Requires-Dist: unidecode==1.3.8
+Requires-Dist: python-docx==1.1.2
+Requires-Dist: psutil
+Requires-Dist: thefuzz==0.22.1
+Requires-Dist: beautifulsoup4==4.12.3
+Requires-Dist: CATEGORIT
+
+UNKNOWN

aait-2.3.12.3/aait.egg-info/PKG-INFO

@@ -0,0 +1,32 @@
+Metadata-Version: 2.1
+Name: aait
+Version: 2.3.12.3
+Summary: Advanced Artificial Intelligence Tools is a package meant to develop and enable advanced AI functionalities in Orange
+Home-page: 
+Author: Orange community
+Author-email: 
+License: UNKNOWN
+Keywords: orange3 add-on
+Platform: UNKNOWN
+License-File: License.txt
+Requires-Dist: torch
+Requires-Dist: sentence-transformers==5.0.0
+Requires-Dist: gpt4all[all]==2.8.2
+Requires-Dist: sacremoses==0.1.1
+Requires-Dist: transformers==4.51.3
+Requires-Dist: sentencepiece==0.2.0
+Requires-Dist: optuna
+Requires-Dist: spacy==3.7.6
+Requires-Dist: markdown
+Requires-Dist: python-multipart
+Requires-Dist: PyMuPDF==1.24.14
+Requires-Dist: chonkie==0.4.1
+Requires-Dist: GPUtil==1.4.0
+Requires-Dist: unidecode==1.3.8
+Requires-Dist: python-docx==1.1.2
+Requires-Dist: psutil
+Requires-Dist: thefuzz==0.22.1
+Requires-Dist: beautifulsoup4==4.12.3
+Requires-Dist: CATEGORIT
+
+UNKNOWN

{aait-2.3.12.2 → aait-2.3.12.3}/orangecontrib/AAIT/widgets/OWFileWithPath.py

@@ -2,9 +2,8 @@ import os
 import sys
 
 import Orange.data
-from AnyQt.QtWidgets import QApplication
+from AnyQt.QtWidgets import QApplication, QPushButton
 from Orange.data import StringVariable
-from Orange.widgets import widget
 from Orange.widgets.utils.signals import Input, Output
 from AnyQt.QtWidgets import QCheckBox
 from Orange.widgets.settings import Setting
@@ -29,7 +28,7 @@ class OWFileWithPath(base_widget.BaseListWidget):
     priority = 1060
 
     # Settings
-    selected_column_name = Setting("content")
+    selected_column_name = Setting("path")
 
     class Inputs:
         filepath = Input("Path", str, auto_summary=False)
@@ -50,14 +49,16 @@ class OWFileWithPath(base_widget.BaseListWidget):
     @Inputs.path_table
     def set_path_table(self, in_path_table):
         self.in_path_table = in_path_table
+        if self.in_path_table is None:
+            self.Outputs.data.send(None)
+            return
         if self.in_path_table is not None:
             self.var_selector.add_variables(self.in_path_table.domain)
             self.var_selector.select_variable_by_name(self.selected_column_name)
-            if "path" in in_path_table.domain:
-                self.filepath = in_path_table[0]["path"].value
-                self.run()
-            else:
-                self.warning("You need a 'path' variable from which the data will be loaded.")
+
+            self.filepath = in_path_table[0][self.selected_column_name].value
+            self.run()
+
         if self.autorun:
             self.run()
 
@@ -66,6 +67,8 @@ class OWFileWithPath(base_widget.BaseListWidget):
         super().__init__()
         # Qt Management
         #uic.loadUi(self.gui, self)
+        self.setFixedWidth(500)
+        self.setFixedHeight(450)
         self.checkbox_interface = self.findChild(QCheckBox, 'checkBox')
         if self.strloadAsString=="False":
             self.checkbox_interface.setChecked(False)
@@ -80,6 +83,9 @@ class OWFileWithPath(base_widget.BaseListWidget):
         self.data = None
         self.autorun = True
         self.post_initialized()
+
+        self.pushButton_run = self.findChild(QPushButton, 'pushButton_send')
+        self.pushButton_run.clicked.connect(self.run)
     def on_checkbox_toggled(self):
         if self.checkbox_interface.isChecked():
             self.strloadAsString="True"

{aait-2.3.12.2 → aait-2.3.12.3}/orangecontrib/AAIT/widgets/OWFindFilesFromDir.py

@@ -1,11 +1,15 @@
-import os
-import sys
 import Orange.data
 from AnyQt.QtWidgets import QApplication, QLineEdit, QCheckBox,QSpinBox, QPushButton
 from Orange.data import StringVariable, Table, Domain
 from Orange.widgets.utils.signals import Input, Output
 from Orange.widgets.settings import Setting
 from datetime import datetime, timedelta
+import openpyxl
+import os
+import sys
+import shutil
+from openpyxl import Workbook, load_workbook
+
 if "site-packages/Orange/widgets" in os.path.dirname(os.path.abspath(__file__)).replace("\\", "/"):
     from Orange.widgets.orangecontrib.AAIT.utils import base_widget
     from Orange.widgets.orangecontrib.AAIT.utils.initialize_from_ini import apply_modification_from_python_file
@@ -30,6 +34,7 @@ class OWFileFromDir(base_widget.BaseListWidget):
     recursive = Setting("False")
     filter_per_date = Setting("False")
     time_filter= Setting("0-0-0-1")
+    identify_sheets = Setting("False")
 
     class Inputs:
         data = Input("Data", Orange.data.Table)
@@ -54,7 +59,7 @@ class OWFileFromDir(base_widget.BaseListWidget):
         super().__init__()
         # Qt Management
         self.setFixedWidth(500)
-        self.setFixedHeight(600)
+        self.setFixedHeight(620)
         if self.is_dash_int_castable(self.time_filter)!=0:
             self.time_filter = "0-0-0-1"
 
@@ -67,8 +72,10 @@ class OWFileFromDir(base_widget.BaseListWidget):
         self.spinBox_day = self.findChild(QSpinBox, 'spinBox_day')
         self.spinBox_hour = self.findChild(QSpinBox, 'spinBox_hour')
         self.spinBox_minute = self.findChild(QSpinBox, 'spinBox_minute')
-        self.spinBox_second= self.findChild(QSpinBox, 'spinBox_second')
-        DD,JJ,SS,MM=self._parse_delta_4_numbers(self.time_filter)
+        self.spinBox_second = self.findChild(QSpinBox, 'spinBox_second')
+        self.checkBox_sheets = self.findChild(QCheckBox, 'checkBox_sheets')
+
+        DD, JJ, SS, MM = self._parse_delta_4_numbers(self.time_filter)
         self.spinBox_day.setValue(int(DD))
         self.spinBox_hour.setValue(int(JJ))
         self.spinBox_minute.setValue(int(SS))
@@ -76,6 +83,7 @@ class OWFileFromDir(base_widget.BaseListWidget):
         # Data Management
         self.folderpath = None
         self.data = None
+        self.temp_dir = None
         self.autorun = True
         self.post_initialized()
 
@@ -95,6 +103,10 @@ class OWFileFromDir(base_widget.BaseListWidget):
         if self.recursive == "True":
             self.comboBox.setChecked(True)
 
+        if self.identify_sheets == "True":
+            self.checkBox_sheets.setChecked(True)
+        else:
+            self.checkBox_sheets.setChecked(False)
 
         self.comboBox.stateChanged.connect(self.on_checkbox_toggled)
         self.checkBox_2.stateChanged.connect(self.on_checkBox_2_toggled)
@@ -102,27 +114,30 @@ class OWFileFromDir(base_widget.BaseListWidget):
         self.spinBox_hour.valueChanged.connect(self.on_value_changed)
         self.spinBox_minute.valueChanged.connect(self.on_value_changed)
         self.spinBox_second.valueChanged.connect(self.on_value_changed)
+        self.checkBox_sheets.stateChanged.connect(self.on_checkBox_sheets_toggled)
 
-        self.pushButton_run =self.findChild(QPushButton, 'pushButton_send')
+        self.pushButton_run = self.findChild(QPushButton, 'pushButton_send')
         self.pushButton_run.clicked.connect(self.run)
 
-
     def on_value_changed(self):
-        self.time_filter = str(self.spinBox_day.value())+"-"+str(self.spinBox_hour.value())+"-"+str(self.spinBox_minute.value())+"-"+str(self.spinBox_second.value())
+        self.time_filter = str(self.spinBox_day.value()) + "-" + str(self.spinBox_hour.value()) + "-" + str(
+            self.spinBox_minute.value()) + "-" + str(self.spinBox_second.value())
+
     def update_parameters(self):
-        self.extension = (self.edit_extension.text() or "").strip() #a jout de la gestion d'une zone vide
+        self.extension = (self.edit_extension.text() or "").strip()  # a jout de la gestion d'une zone vide
         if self.folderpath is not None:
             self.run()
 
-    def on_checkbox_toggled(self,state):
+    def on_checkbox_toggled(self, state):
         self.recursive = "True"
-        if state==0:
+        if state == 0:
             self.recursive = "False"
         if self.folderpath is not None:
             self.run()
-    def on_checkBox_2_toggled(self,state):
+
+    def on_checkBox_2_toggled(self, state):
         self.filter_per_date = "True"
-        if state==0:
+        if state == 0:
             self.filter_per_date = "False"
 
         if self.filter_per_date == "False":
@@ -136,16 +151,79 @@ class OWFileFromDir(base_widget.BaseListWidget):
             self.spinBox_minute.setVisible(True)
             self.spinBox_second.setVisible(True)
 
+        if self.folderpath is not None:
+            self.run()
 
-
-
+    def on_checkBox_sheets_toggled(self, state):
+        self.identify_sheets = "True"
+        if state == 0:
+            self.identify_sheets = "False"
         if self.folderpath is not None:
             self.run()
 
+    def split_excel_sheets(self, filepath):
+        """
+        Découpe un fichier Excel.
+        Crée un dossier [NomFichier]_sheets dans le même répertoire que le fichier.
+        Y dépose un fichier .xlsx par feuille.
+        """
+        new_paths = []
+        try:
+            source_dir = os.path.dirname(filepath)
+            filename_no_ext = os.path.splitext(os.path.basename(filepath))[0]
+
+            # Nom du dossier : ex "MonFichier_sheets"
+            output_folder_name = f"{filename_no_ext}_sheets"
+            output_folder_path = os.path.join(source_dir, output_folder_name)
+
+            # Création du dossier (exist_ok=True évite de planter s'il existe déjà)
+            os.makedirs(output_folder_path, exist_ok=True)
+
+            # 2. Lire le fichier source avec Openpyxl
+            wb = load_workbook(filepath, data_only=True)
+
+            for sheet_name in wb.sheetnames:
+                ws = wb[sheet_name]
+
+                safe_sheet = "".join(c if c.isalnum() else "_" for c in sheet_name)
+                target_filename = f"{safe_sheet}.xlsx"
+                target_path = os.path.join(output_folder_path, target_filename).replace("\\", "/")
+
+                new_wb = Workbook()
+                new_ws = new_wb.active
+                new_ws.title = sheet_name
+
+                for row in ws.iter_rows():
+                    new_ws.append([
+                        "" if cell.value is None else str(cell.value)
+                        for cell in row
+                    ])
+
+                new_wb.save(target_path)
+                new_paths.append(target_path)
+
+            wb.close()
+
+        except Exception as e:
+            # En cas d'échec (ex: fichier ouvert, permissions...), on log et on renvoie l'original
+            print(f"Erreur lors de l'extraction des feuilles pour {filepath} : {e}")
+            return [filepath]
+
+        return new_paths
 
     def find_files(self):
         files_data = []
-        suffixes = self.parse_extensions()  # si pas de d'extensions, elles apparaitront toutes
+        suffixes = self.parse_extensions()
+        excel_suffixes = (".xlsx", ".xlsm", ".xls")
+
+        # Nettoyage du dossier temp précédent si nouvelle recherche
+        # (Note : ce bloc est conservé tel quel, même si on utilise maintenant des dossiers locaux)
+        if self.identify_sheets == "True" and self.temp_dir and os.path.exists(self.temp_dir):
+            try:
+                shutil.rmtree(self.temp_dir)
+                self.temp_dir = None
+            except:
+                pass
 
         for i in range(len(self.folderpath)):
             base = self.folderpath[i]
@@ -160,8 +238,42 @@ class OWFileFromDir(base_widget.BaseListWidget):
             for root, _, files in traversal:
                 for file in files:
                     name = file.lower()
+                    full_path = os.path.join(root, file).replace("\\", "/")
+
                     if suffixes is None or name.endswith(suffixes):
-                        files_data.append([os.path.join(root, file).replace("\\", "/")])
+                        is_excel = any(name.endswith(ext) for ext in excel_suffixes)
+
+                        if self.identify_sheets == "True" and is_excel:
+                            filename_no_ext = os.path.splitext(file)[0]
+                            expected_folder_name = f"{filename_no_ext}_sheets"
+                            expected_folder_path = os.path.join(root, expected_folder_name)
+
+                            if os.path.exists(expected_folder_path) and os.path.isdir(expected_folder_path):
+                                found_sheets = False
+                                try:
+                                    existing_files = [
+                                        os.path.join(expected_folder_path, f).replace("\\", "/")
+                                        for f in os.listdir(expected_folder_path)
+                                        if f.endswith(".xlsx") and not f.startswith("~$")
+                                    ]
+                                    if existing_files:
+                                        for ex_path in existing_files:
+                                            files_data.append([ex_path])
+                                        found_sheets = True
+                                except OSError:
+                                    pass
+                                # Si le dossier existait mais était vide
+                                if not found_sheets:
+                                    split_files = self.split_excel_sheets(full_path)
+                                    for split_path in split_files:
+                                        files_data.append([split_path])
+                            # 3. Le dossier n'existe pas, on lance le découpage
+                            else:
+                                split_files = self.split_excel_sheets(full_path)
+                                for split_path in split_files:
+                                    files_data.append([split_path])
+                        else:
+                            files_data.append([full_path])
         return files_data
 
     def parse_extensions(self):
@@ -330,14 +442,36 @@ class OWFileFromDir(base_widget.BaseListWidget):
     def post_initialized(self):
         pass
 
+    def get_sheet_names_from_file(self, filepath: str):
+        """
+        Méthode simplifiée pour récupérer la liste des noms des feuilles
+        en utilisant directement les bibliothèques sous-jacentes (openpyxl/xlrd).
+        """
+        # Récupérer l'extension en minuscules
+        extension = os.path.splitext(filepath)[1].lower()
+        sheets = []
+
+        try:
+            if extension in ('.xlsx', '.xlsm'):
+                workbook = openpyxl.load_workbook(filepath, read_only=True, data_only=False)
+                sheets = workbook.sheetnames
+                workbook.close()
+            else:
+                print(f"INFO: Extension {extension} non supportée pour l'extraction de feuilles.")
+
+        except ImportError as e:
+            self.warning(
+                f"Librairie manquante pour lire les feuilles : {e}. Assurez-vous d'avoir 'openpyxl'")
+        except Exception as e:
+            self.error(f"Impossible de lire les feuilles du fichier {os.path.basename(filepath)} : {e}")
+        return sheets
+
 if __name__ == "__main__":
     app = QApplication(sys.argv)
-    #data = Orange.data.Table(r"C:\Users\max83\Desktop\data.tab")
     my_widget = OWFileFromDir()
     my_widget.show()
-    #my_widget.set_path_table(data)
 
     if hasattr(app, "exec"):
         app.exec()
     else:
-        app.exec_()
+        app.exec_()

{aait-2.3.12.2 → aait-2.3.12.3}/orangecontrib/AAIT/widgets/designer/owfindfilesfromdir.ui

@@ -7,7 +7,7 @@
     <x>0</x>
     <y>0</y>
     <width>478</width>
-    <height>563</height>
+    <height>603</height>
    </rect>
   </property>
   <property name="windowTitle">
@@ -19,7 +19,7 @@
      <x>19</x>
      <y>30</y>
      <width>441</width>
-     <height>281</height>
+     <height>321</height>
     </rect>
    </property>
    <property name="title">
@@ -61,11 +61,25 @@
      <string>Allow multiple extensions separated with comma (.jpg,.png)</string>
     </property>
    </widget>
+
+   <widget class="QCheckBox" name="checkBox_sheets">
+    <property name="geometry">
+     <rect>
+      <x>20</x>
+      <y>160</y>
+      <width>411</width>
+      <height>16</height>
+     </rect>
+    </property>
+    <property name="text">
+     <string>Also identify sheets in Excel files (.xlsx, .xlsm)</string>
+    </property>
+   </widget>
    <widget class="QCheckBox" name="checkBox_2">
     <property name="geometry">
      <rect>
       <x>20</x>
-      <y>180</y>
+      <y>200</y>
       <width>411</width>
       <height>16</height>
      </rect>
@@ -78,7 +92,7 @@
     <property name="geometry">
      <rect>
       <x>30</x>
-      <y>220</y>
+      <y>240</y>
       <width>71</width>
       <height>22</height>
      </rect>
@@ -91,7 +105,7 @@
     <property name="geometry">
      <rect>
       <x>120</x>
-      <y>220</y>
+      <y>240</y>
       <width>71</width>
       <height>22</height>
      </rect>
@@ -104,7 +118,7 @@
     <property name="geometry">
      <rect>
       <x>210</x>
-      <y>220</y>
+      <y>240</y>
       <width>71</width>
       <height>22</height>
      </rect>
@@ -117,7 +131,7 @@
     <property name="geometry">
      <rect>
       <x>300</x>
-      <y>220</y>
+      <y>240</y>
       <width>71</width>
       <height>22</height>
      </rect>
@@ -137,7 +151,7 @@
    <property name="geometry">
     <rect>
      <x>20</x>
-     <y>330</y>
+     <y>370</y>
      <width>431</width>
      <height>161</height>
     </rect>
@@ -188,7 +202,7 @@
    <property name="geometry">
     <rect>
      <x>149</x>
-     <y>514</y>
+     <y>554</y>
      <width>301</width>
      <height>31</height>
     </rect>
@@ -204,7 +218,7 @@
    <property name="geometry">
     <rect>
      <x>20</x>
-     <y>520</y>
+     <y>560</y>
      <width>131</width>
      <height>16</height>
     </rect>
@@ -216,4 +230,4 @@
  </widget>
  <resources/>
  <connections/>
-</ui>
+</ui>

{aait-2.3.12.2 → aait-2.3.12.3}/setup.py

@@ -4,7 +4,7 @@ from setuptools import setup, find_packages
 NAME = "aait"
 
 # Version du package PyPI
-VERSION = "2.3.12.2"  # la version doit être supérieure à la précédente sinon la publication sera refusée
+VERSION = "2.3.12.3"  # la version doit être supérieure à la précédente sinon la publication sera refusée
 
 # Facultatif / Adaptable à souhait
 AUTHOR = "Orange community"

aait-2.3.12.2/PKG-INFO

@@ -1,9 +0,0 @@
-Metadata-Version: 2.1
-Name: aait
-Version: 2.3.12.2
-Summary: Advanced Artificial Intelligence Tools is a package meant to develop and enable advanced AI functionalities in Orange
-Home-page: 
-Author: Orange community
-Author-email: 
-Keywords: orange3 add-on
-License-File: License.txt

aait-2.3.12.2/aait.egg-info/PKG-INFO

@@ -1,9 +0,0 @@
-Metadata-Version: 2.1
-Name: aait
-Version: 2.3.12.2
-Summary: Advanced Artificial Intelligence Tools is a package meant to develop and enable advanced AI functionalities in Orange
-Home-page: 
-Author: Orange community
-Author-email: 
-Keywords: orange3 add-on
-License-File: License.txt