io4it 3.0.2.1__tar.gz → 3.0.3__tar.gz

{io4it-3.0.2.1 → io4it-3.0.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: io4it
-Version: 3.0.2.1
+Version: 3.0.3
 Home-page: 
 Author: 
 Author-email: 

{io4it-3.0.2.1 → io4it-3.0.3}/io4it.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: io4it
-Version: 3.0.2.1
+Version: 3.0.3
 Home-page: 
 Author: 
 Author-email: 

{io4it-3.0.2.1 → io4it-3.0.3}/io4it.egg-info/SOURCES.txt

@@ -1,3 +1,4 @@
+setup.cfg
 setup.py
 io4it.egg-info/PKG-INFO
 io4it.egg-info/SOURCES.txt

{io4it-3.0.2.1 → io4it-3.0.3}/orangecontrib/IO4IT/utils/utils_md.py

@@ -1,6 +1,7 @@
 import re, hashlib, secrets, urllib.parse, logging, numpy as np
 from pathlib import Path
 import fitz  # PyMuPDF
+import win32com.client
 
 try:
     import easyocr
@@ -88,7 +89,6 @@ def _make_powerpoint_invisible(ppt):
         pass
 
 def convert_doc_to_docx(src: Path, out_dir: Path) -> Path:
-    import win32com.client
     out_dir.mkdir(parents=True, exist_ok=True)
     if src.suffix.lower() != ".doc":
         return src if src.suffix.lower() == ".docx" else src
@@ -108,7 +108,7 @@ def convert_doc_to_docx(src: Path, out_dir: Path) -> Path:
         word.Quit()
 
 def convert_ppt_to_pptx(src: Path, out_dir: Path) -> Path:
-    import win32com.client
+
     out_dir.mkdir(parents=True, exist_ok=True)
     if src.suffix.lower() != ".ppt":
         return src if src.suffix.lower() == ".pptx" else src
@@ -128,7 +128,6 @@ def convert_ppt_to_pptx(src: Path, out_dir: Path) -> Path:
         ppt.Quit()
 
 def docx_to_pdf(src_docx: Path, out_dir: Path) -> Path:
-    import win32com.client
     out_dir.mkdir(parents=True, exist_ok=True)
     pdf_path = out_dir / (src_docx.stem + ".pdf")
     if pdf_path.exists() and pdf_path.stat().st_size > 0:
@@ -149,7 +148,6 @@ def docx_to_pdf(src_docx: Path, out_dir: Path) -> Path:
         word.Quit()
 
 def pptx_to_pdf(src_pptx: Path, out_dir: Path) -> Path:
-    import win32com.client
     out_dir.mkdir(parents=True, exist_ok=True)
     pdf_path = out_dir / (src_pptx.stem + ".pdf")
     if pdf_path.exists() and pdf_path.stat().st_size > 0:
@@ -165,3 +163,10 @@ def pptx_to_pdf(src_pptx: Path, out_dir: Path) -> Path:
         return pdf_path
     finally:
         ppt.Quit()
+
+def is_word_installed():
+    try:
+        win32com.client.Dispatch("Word.Application")
+        return True
+    except Exception:
+        return False

{io4it-3.0.2.1 → io4it-3.0.3}/orangecontrib/IO4IT/widgets/OWExportMarkdown.py

@@ -19,8 +19,10 @@ from docx2pdf import convert
 # Chargement UI
 if "site-packages/Orange/widgets" in os.path.dirname(os.path.abspath(__file__)).replace("\\", "/"):
     from Orange.widgets.orangecontrib.AAIT.utils.import_uic import uic
+    from Orange.widgets.orangecontrib.IO4IT.utils import utils_md
 else:
     from orangecontrib.AAIT.utils.import_uic import uic
+    from orangecontrib.IO4IT.utils import utils_md
 
 
 class OWExportMarkdown(widget.OWWidget):
@@ -164,32 +166,30 @@ class OWExportMarkdown(widget.OWWidget):
                 tmp_md = tmp.name
 
             try:
-                # DOCX
-                pypandoc.convert_file(tmp_md, to="docx", format="gfm-yaml_metadata_block", outputfile=docx_out)
-                self.ajouter_en_tete_pied_docx(
-                    docx_out,
-                    "Rapport - Orange AI",
-                    "Page générée automatiquement - Ne pas diffuser"
-                )
-
-                # PPTX
-                pypandoc.convert_file(tmp_md, to="pptx", format="gfm-yaml_metadata_block", outputfile=pptx_out)
-                self.ajouter_entete_pied_pptx(
-                    pptx_out,
-                    "Orange AI – Présentation",
-                    "Page générée automatiquement"
-                )
-
-                # PDF (docx -> pdf) avec chemin de sortie exact
+                if utils_md.is_word_installed():
+                    pypandoc.convert_file(tmp_md, to="docx", format="gfm-yaml_metadata_block", outputfile=docx_out)
+                    self.ajouter_en_tete_pied_docx(
+                        docx_out,
+                        "Rapport - Orange AI",
+                        "Page générée automatiquement - Ne pas diffuser"
+                    )
+
+                    # PPTX
+                    pypandoc.convert_file(tmp_md, to="pptx", format="gfm-yaml_metadata_block", outputfile=pptx_out)
+                    self.ajouter_entete_pied_pptx(
+                        pptx_out,
+                        "Orange AI – Présentation",
+                        "Page générée automatiquement"
+                    )
+                else:
+                    raise Exception("Word non détecté")
+
+            except Exception:
                 try:
-                    convert(docx_out, pdf_out)
+                    pypandoc.convert_file(tmp_md, to="pdf", outputfile=pdf_out)
                 except Exception:
-                    # fallback pandoc->pdf (si LaTeX dispo)
-                    try:
-                        pypandoc.convert_file(tmp_md, to="pdf", outputfile=pdf_out)
-                    except Exception:
-                        self.error(f"Échec conversion PDF pour la ligne {i+1}.")
-                        pdf_out = ""
+                    self.error(f"Échec conversion PDF pour la ligne {i + 1}.")
+                    pdf_out = ""
             finally:
                 try:
                     os.remove(tmp_md)

{io4it-3.0.2.1 → io4it-3.0.3}/orangecontrib/IO4IT/widgets/OWPdfType.py

@@ -1,12 +1,13 @@
 import os
 import sys
 from pathlib import Path
-
-from AnyQt.QtWidgets import QApplication
 from AnyQt.QtCore import pyqtSignal
 from Orange.data import Domain, StringVariable, Table, DiscreteVariable
-from Orange.widgets import widget
+
+import fitz  # PyMuPDF
+from AnyQt.QtWidgets import QApplication, QCheckBox, QPushButton
 from Orange.widgets.utils.signals import Input, Output
+from Orange.widgets.settings import Setting
 
 # --- Ajout pour l'écriture Excel ---
 from openpyxl import Workbook
@@ -15,14 +16,14 @@ from openpyxl import Workbook
 if "site-packages/Orange/widgets" in os.path.dirname(os.path.abspath(__file__)).replace("\\", "/"):
     from Orange.widgets.orangecontrib.IO4IT.utils import utils_md
     from Orange.widgets.orangecontrib.AAIT.utils.thread_management import Thread
-    from Orange.widgets.orangecontrib.AAIT.utils.import_uic import uic
+    from Orange.widgets.orangecontrib.AAIT.utils import base_widget
 else:
     from orangecontrib.IO4IT.utils import utils_md
     from orangecontrib.AAIT.utils.thread_management import Thread
-    from orangecontrib.AAIT.utils.import_uic import uic
+    from orangecontrib.AAIT.utils import base_widget
 
 
-class OWPdfType(widget.OWWidget):
+class OWPdfType(base_widget.BaseListWidget):
     name = "PDF Type"
     description = "Checks if a PDF is text-based or image-based"
     category = "AAIT - TOOLBOX"
@@ -32,6 +33,10 @@ class OWPdfType(widget.OWWidget):
     gui = os.path.join(os.path.dirname(os.path.abspath(__file__)), "designer/owpdftype.ui")
     want_control_area = False
     priority = 1002
+    recursive = Setting("False")
+
+    # Settings
+    selected_column_name = Setting("content")
 
     MAX_PAGES_TO_CHECK = 5
 
@@ -49,8 +54,8 @@ class OWPdfType(widget.OWWidget):
     def __init__(self):
         super().__init__()
         self.setFixedWidth(470)
-        self.setFixedHeight(300)
-        uic.loadUi(self.gui, self)
+        self.setFixedHeight(470)
+        #uic.loadUi(self.gui, self)
 
         self.data = None
         self.thread = None
@@ -59,9 +64,22 @@ class OWPdfType(widget.OWWidget):
         self.processed_statuses = []
         self.post_initialized()
 
+        self.comboBox = self.findChild(QCheckBox, 'checkBox_send')
+        self.pushButton_run = self.findChild(QPushButton, 'pushButton_send')
+
+
+        if self.recursive == "True":
+            self.comboBox.setChecked(True)
+
+        self.comboBox.stateChanged.connect(self.on_checkbox_toggled)
+        self.pushButton_run.clicked.connect(self.run)
+
     @Inputs.data
     def set_data(self, in_data: Table | None):
         self.data = in_data
+        if self.data:
+            self.var_selector.add_variables(self.data.domain)
+            self.var_selector.select_variable_by_name(self.selected_column_name)
         if self.autorun:
             self.run()
 
@@ -76,14 +94,14 @@ class OWPdfType(widget.OWWidget):
             return
 
         self.error("")
-        try:
-            self.data.domain["file_path"]
-        except KeyError:
-            self.error("You need a 'file_path' column in input data.")
+
+        # Verification of in_data
+        if not self.selected_column_name in self.data.domain:
+            self.warning(f'Previously selected column "{self.selected_column_name}" does not exist in your data.')
             return
 
-        if type(self.data.domain["file_path"]).__name__ != 'StringVariable':
-            self.error("'file_path' column needs to be a Text.")
+        if not isinstance(self.data.domain[self.selected_column_name], StringVariable):
+            self.error('You must select a text variable.')
             return
 
         self.progressBarInit()
@@ -99,17 +117,14 @@ class OWPdfType(widget.OWWidget):
         self.thread.finish.connect(self.handle_finish)
         self.thread.start()
 
-    def _process_pdfs(self, in_data: Table, progress_callback: callable, status_callback: callable) -> tuple[
-        Table | None, Table | None]:
-
-        # Extraction des chemins de fichiers avant de commencer le traitement
-        paths = [str(x) for x in in_data.get_column("file_path")]
+    def _process_pdfs(self, in_data: Table, progress_callback: callable, status_callback: callable):
+        paths = [str(x) for x in in_data.get_column(self.selected_column_name)]
 
+        # Dossier Excel
         excel_output_dir = Path.cwd() / "pdf_check_results"
         if paths:
             first_file_path = Path(paths[0])
             excel_output_dir = first_file_path.parent / "pdf_check_results"
-
         excel_output_dir.mkdir(parents=True, exist_ok=True)
 
         base_name = "pdf_check_results"
@@ -122,7 +137,7 @@ class OWPdfType(widget.OWWidget):
         wb = Workbook()
         ws = wb.active
         ws.title = "PDF Check Results"
-        headers = ["file_path", "status", "details"]
+        headers = [self.selected_column_name, "status", "details"]
         ws.append(headers)
 
         text_indices = []
@@ -142,6 +157,28 @@ class OWPdfType(widget.OWWidget):
                 wb.save(excel_path)
                 continue
 
+            # 🔒 --- Vérification verrouillage via PyMuPDF ---
+            try:
+                doc = fitz.open(str(fp))
+
+                if doc.is_encrypted:
+                    result_row[1] = "ko"
+                    result_row[2] = "Locked/Encrypted PDF"
+                    status_callback(result_row)
+                    ws.append(result_row)
+                    wb.save(excel_path)
+                    continue
+
+            except Exception as e:
+                # PyMuPDF lève une exception si le PDF est trop protégé
+                result_row[1] = "ko"
+                result_row[2] = f"Cannot open PDF (possibly locked): {str(e)}"
+                status_callback(result_row)
+                ws.append(result_row)
+                wb.save(excel_path)
+                continue
+
+            # --- Détection texte / image via utils_md ---
             try:
                 is_text = utils_md.is_pdf_text_based(fp)
                 if is_text:
@@ -165,15 +202,8 @@ class OWPdfType(widget.OWWidget):
 
         progress_callback(100)
 
-        if not text_indices:
-            text_table = None
-        else:
-            text_table = in_data[text_indices]
-
-        if not image_indices:
-            image_table = None
-        else:
-            image_table = in_data[image_indices]
+        text_table = in_data[text_indices] if text_indices else None
+        image_table = in_data[image_indices] if image_indices else None
 
         return text_table, image_table
 
@@ -191,7 +221,7 @@ class OWPdfType(widget.OWWidget):
         status_domain = Domain(
             [],  # The variables list should be empty
             metas=[
-                StringVariable("file_path"),
+                StringVariable(self.selected_column_name),
                 DiscreteVariable("status", values=["ok", "ko"]),
                 StringVariable("details")
             ]
@@ -216,6 +246,13 @@ class OWPdfType(widget.OWWidget):
     def post_initialized(self):
         pass
 
+    def on_checkbox_toggled(self,state):
+        self.recursive = "True"
+        if state==0:
+            self.recursive = "False"
+        if self.data is not None:
+            self.run()
+
 
 if __name__ == "__main__":
     app = QApplication(sys.argv)

io4it-3.0.3/orangecontrib/IO4IT/widgets/designer/owpdftype.ui

@@ -0,0 +1,131 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<ui version="4.0">
+ <class>Form</class>
+ <widget class="QWidget" name="Form">
+  <property name="geometry">
+   <rect>
+    <x>0</x>
+    <y>0</y>
+    <width>474</width>
+    <height>460</height>
+   </rect>
+  </property>
+  <property name="windowTitle">
+   <string>Form</string>
+  </property>
+  <widget class="QGroupBox" name="groupBox">
+   <property name="geometry">
+    <rect>
+     <x>10</x>
+     <y>30</y>
+     <width>451</width>
+     <height>181</height>
+    </rect>
+   </property>
+   <property name="title">
+    <string>Parameters</string>
+   </property>
+   <widget class="QLabel" name="Description">
+    <property name="geometry">
+     <rect>
+      <x>10</x>
+      <y>40</y>
+      <width>431</width>
+      <height>101</height>
+     </rect>
+    </property>
+    <property name="text">
+     <string>This widget analyze PDF file paths and determines if each PDF is text-based or image-based, adding a &quot;type&quot; meta column to the output data.</string>
+    </property>
+    <property name="textFormat">
+     <enum>Qt::AutoText</enum>
+    </property>
+    <property name="alignment">
+     <set>Qt::AlignLeading|Qt::AlignLeft|Qt::AlignVCenter</set>
+    </property>
+    <property name="wordWrap">
+     <bool>true</bool>
+    </property>
+   </widget>
+  </widget>
+  <widget class="QPushButton" name="pushButton_send">
+   <property name="enabled">
+    <bool>false</bool>
+   </property>
+   <property name="geometry">
+    <rect>
+     <x>170</x>
+     <y>410</y>
+     <width>291</width>
+     <height>31</height>
+    </rect>
+   </property>
+   <property name="text">
+    <string>Run</string>
+   </property>
+  </widget>
+  <widget class="QCheckBox" name="checkBox_send">
+   <property name="enabled">
+    <bool>false</bool>
+   </property>
+   <property name="geometry">
+    <rect>
+     <x>20</x>
+     <y>410</y>
+     <width>131</width>
+     <height>16</height>
+    </rect>
+   </property>
+   <property name="text">
+    <string>Auto send data</string>
+   </property>
+  </widget>
+  <widget class="QGroupBox" name="groupBox_3">
+   <property name="geometry">
+    <rect>
+     <x>10</x>
+     <y>230</y>
+     <width>451</width>
+     <height>161</height>
+    </rect>
+   </property>
+   <property name="title">
+    <string>Column selection</string>
+   </property>
+   <widget class="QLabel" name="Description_4">
+    <property name="geometry">
+     <rect>
+      <x>10</x>
+      <y>30</y>
+      <width>411</width>
+      <height>61</height>
+     </rect>
+    </property>
+    <property name="text">
+     <string/>
+    </property>
+    <property name="textFormat">
+     <enum>Qt::AutoText</enum>
+    </property>
+    <property name="alignment">
+     <set>Qt::AlignLeading|Qt::AlignLeft|Qt::AlignVCenter</set>
+    </property>
+    <property name="wordWrap">
+     <bool>true</bool>
+    </property>
+   </widget>
+   <widget class="QWidget" name="placeholder" native="true">
+    <property name="geometry">
+     <rect>
+      <x>0</x>
+      <y>20</y>
+      <width>431</width>
+      <height>141</height>
+     </rect>
+    </property>
+   </widget>
+  </widget>
+ </widget>
+ <resources/>
+ <connections/>
+</ui>

{io4it-3.0.2.1 → io4it-3.0.3}/setup.py

@@ -2,7 +2,7 @@ from setuptools import setup, find_packages
 
 # Configuration
 NAME = "io4it"
-VERSION = "3.0.2.1"
+VERSION = "3.0.3"
 
 INSTALL_REQUIRES = [
     "torchvision==0.23.0",

io4it-3.0.2.1/orangecontrib/IO4IT/widgets/designer/owpdftype.ui

@@ -1,86 +0,0 @@
-<?xml version="1.0" encoding="UTF-8"?>
-<ui version="4.0">
- <class>Form</class>
- <widget class="QWidget" name="Form">
-  <property name="geometry">
-   <rect>
-    <x>0</x>
-    <y>0</y>
-    <width>470</width>
-    <height>305</height>
-   </rect>
-  </property>
-  <property name="windowTitle">
-   <string>Form</string>
-  </property>
-  <widget class="QGroupBox" name="groupBox">
-   <property name="geometry">
-    <rect>
-     <x>10</x>
-     <y>30</y>
-     <width>451</width>
-     <height>181</height>
-    </rect>
-   </property>
-   <property name="title">
-    <string>Parameters</string>
-   </property>
-   <widget class="QLabel" name="Description">
-    <property name="geometry">
-     <rect>
-      <x>10</x>
-      <y>40</y>
-      <width>431</width>
-      <height>101</height>
-     </rect>
-    </property>
-    <property name="text">
-     <string>This widget analyzes a column named &quot;file_path&quot; containing PDF file paths and determines if each PDF is text-based or image-based, adding a &quot;type&quot; meta column to the output data.</string>
-    </property>
-    <property name="textFormat">
-     <enum>Qt::AutoText</enum>
-    </property>
-    <property name="alignment">
-     <set>Qt::AlignLeading|Qt::AlignLeft|Qt::AlignVCenter</set>
-    </property>
-    <property name="wordWrap">
-     <bool>true</bool>
-    </property>
-   </widget>
-  </widget>
-  <widget class="QPushButton" name="pushButton_send">
-   <property name="enabled">
-    <bool>false</bool>
-   </property>
-   <property name="geometry">
-    <rect>
-     <x>170</x>
-     <y>244</y>
-     <width>291</width>
-     <height>31</height>
-    </rect>
-   </property>
-   <property name="text">
-    <string>Run</string>
-   </property>
-  </widget>
-  <widget class="QCheckBox" name="checkBox_send">
-   <property name="enabled">
-    <bool>false</bool>
-   </property>
-   <property name="geometry">
-    <rect>
-     <x>10</x>
-     <y>250</y>
-     <width>131</width>
-     <height>16</height>
-    </rect>
-   </property>
-   <property name="text">
-    <string>Auto send data</string>
-   </property>
-  </widget>
- </widget>
- <resources/>
- <connections/>
-</ui>