IO4IT 0.0.0.1__tar.gz

IO4IT-0.0.0.1/IO4IT.egg-info/PKG-INFO

@@ -0,0 +1,8 @@
+Metadata-Version: 2.1
+Name: IO4IT
+Version: 0.0.0.1
+Home-page: 
+Author: 
+Author-email: 
+Keywords: orange3 add-on
+License-File: License.txt

IO4IT-0.0.0.1/IO4IT.egg-info/SOURCES.txt

@@ -0,0 +1,23 @@
+License.txt
+setup.cfg
+setup.py
+IO4IT.egg-info/PKG-INFO
+IO4IT.egg-info/SOURCES.txt
+IO4IT.egg-info/dependency_links.txt
+IO4IT.egg-info/entry_points.txt
+IO4IT.egg-info/namespace_packages.txt
+IO4IT.egg-info/requires.txt
+IO4IT.egg-info/top_level.txt
+orangecontrib/IOIT/__init__.py
+orangecontrib/IOIT/widgets/OWS3downloader.py
+orangecontrib/IOIT/widgets/OWS3list.py
+orangecontrib/IOIT/widgets/OWwordpdf2docx.py
+orangecontrib/IOIT/widgets/__init__.py
+orangecontrib/IOIT/widgets/designer/__init__.py
+orangecontrib/IOIT/widgets/designer/wordpdf2docx.ui
+orangecontrib/IOIT/widgets/icons/__init__.py
+orangecontrib/IOIT/widgets/icons/category.svg
+orangecontrib/IOIT/widgets/icons/list_aws.png
+orangecontrib/IOIT/widgets/icons/upload.png
+orangecontrib/IOIT/widgets/ocr_function/__init__.py
+orangecontrib/IOIT/widgets/ocr_function/word_converter.py

IO4IT-0.0.0.1/IO4IT.egg-info/dependency_links.txt

@@ -0,0 +1 @@
+

IO4IT-0.0.0.1/IO4IT.egg-info/entry_points.txt

@@ -0,0 +1,2 @@
+[orange.widgets]
+Advanced Artificial Intelligence Tools = orangecontrib.AAIT.widgets

IO4IT-0.0.0.1/IO4IT.egg-info/namespace_packages.txt

@@ -0,0 +1 @@
+orangecontrib

IO4IT-0.0.0.1/IO4IT.egg-info/requires.txt

@@ -0,0 +1,2 @@
+boto3
+fastapi

IO4IT-0.0.0.1/IO4IT.egg-info/top_level.txt

@@ -0,0 +1 @@
+orangecontrib

IO4IT-0.0.0.1/License.txt

@@ -0,0 +1,6 @@
+THIS SOFTWARE IS PROVIDED "AS IS", WITHOUT ANY WARRANTY WHATSOEVER.
+
+If you use or redistribute this software, you are permitted to do so
+under the terms of GNU [GPL-3.0]+ license.
+
+[GPL-3.0]: https://www.gnu.org/licenses/gpl-3.0.en.html

IO4IT-0.0.0.1/PKG-INFO

@@ -0,0 +1,8 @@
+Metadata-Version: 2.1
+Name: IO4IT
+Version: 0.0.0.1
+Home-page: 
+Author: 
+Author-email: 
+Keywords: orange3 add-on
+License-File: License.txt

IO4IT-0.0.0.1/orangecontrib/IOIT/widgets/OWS3downloader.py

@@ -0,0 +1,94 @@
+import os
+import boto3
+from Orange.widgets.widget import OWWidget, Input
+from Orange.widgets.settings import Setting
+from Orange.widgets import gui
+from AnyQt.QtWidgets import QLineEdit, QFileDialog,QApplication
+from Orange.data import Table, Domain, StringVariable, ContinuousVariable
+
+class OWS3FileDownloader(OWWidget):
+    name = "S3 File Downloader"
+    description = "Télécharge les fichiers listés depuis S3 vers un répertoire local."
+    icon = "icons/download.png"
+    if "site-packages/Orange/widgets" in os.path.dirname(os.path.abspath(__file__)).replace("\\", "/"):
+        icon = "icons_dev/download.png"
+    priority = 20
+
+    # Paramètres utilisateur
+    access_key = Setting("")
+    secret_key = Setting("")
+    bucket_name = Setting("")
+    download_path = Setting("")
+
+    class Inputs:
+        data = Input("Data", Table)
+
+    def __init__(self):
+        super().__init__()
+        self.data = None
+
+        # Interface utilisateur
+        self.access_key_input = QLineEdit(self)
+        self.access_key_input.setPlaceholderText("AWS Access Key")
+        self.access_key_input.setText(self.access_key)
+        self.access_key_input.editingFinished.connect(self.update_settings)
+        gui.widgetBox(self.controlArea, orientation='vertical').layout().addWidget(self.access_key_input)
+
+        self.secret_key_input = QLineEdit(self)
+        self.secret_key_input.setPlaceholderText("AWS Secret Key")
+        self.secret_key_input.setText(self.secret_key)
+        self.secret_key_input.editingFinished.connect(self.update_settings)
+        gui.widgetBox(self.controlArea, orientation='vertical').layout().addWidget(self.secret_key_input)
+
+        self.bucket_input = QLineEdit(self)
+        self.bucket_input.setPlaceholderText("Bucket S3")
+        self.bucket_input.setText(self.bucket_name)
+        self.bucket_input.editingFinished.connect(self.update_settings)
+        gui.widgetBox(self.controlArea, orientation='vertical').layout().addWidget(self.bucket_input)
+
+        self.path_button = gui.button(self.controlArea, self, "Choisir un répertoire", callback=self.select_directory)
+
+        gui.button(self.controlArea, self, "Télécharger les fichiers", callback=self.download_files)
+
+    def update_settings(self):
+        self.access_key = self.access_key_input.text()
+        self.secret_key = self.secret_key_input.text()
+        self.bucket_name = self.bucket_input.text()
+
+    def select_directory(self):
+        directory = QFileDialog.getExistingDirectory(self, "Sélectionner un répertoire")
+        if directory:
+            self.download_path = directory
+
+    @Inputs.data
+    def set_data(self, data):
+        self.data = data
+
+    def download_files(self):
+        if self.data is None:
+            self.error("Aucune donnée reçue.")
+            return
+
+        if not self.download_path:
+            self.error("Veuillez choisir un répertoire de téléchargement.")
+            return
+
+        try:
+            session = boto3.Session(
+                aws_access_key_id=self.access_key,
+                aws_secret_access_key=self.secret_key
+            )
+            s3 = session.client("s3")
+            for elem in self.data:
+                file_key = elem["Nom"].value
+                local_file_path = os.path.join(self.download_path, file_key)
+
+                os.makedirs(os.path.dirname(local_file_path), exist_ok=True)
+                s3.download_file(self.bucket_name, file_key, local_file_path)
+            self.information("Téléchargement terminé !")
+
+        except Exception as e:
+            print(e)
+            self.error(str(e))
+
+

IO4IT-0.0.0.1/orangecontrib/IOIT/widgets/OWS3list.py

@@ -0,0 +1,107 @@
+import boto3
+from Orange.widgets.widget import OWWidget, Input, Output
+from Orange.widgets.settings import Setting
+from Orange.widgets import gui
+from AnyQt.QtWidgets import QLineEdit, QApplication
+from Orange.data import Table, Domain, StringVariable, ContinuousVariable
+import os
+
+class OWS3FileLister(OWWidget):
+    name = "S3 File Lister"
+    description = "Liste les fichiers d'un bucket S3 et affiche leurs détails."
+    icon = "icons/list_aws.png"
+    if "site-packages/Orange/widgets" in os.path.dirname(os.path.abspath(__file__)).replace("\\", "/"):
+        icon = "icons_dev/list_aws.png"
+    priority = 10
+
+    # Paramètres utilisateur
+    access_key = Setting("")
+    secret_key = Setting("")
+    bucket_name = Setting("")
+    allowed_extensions = Setting("*")
+    folder_name = Setting("")
+
+    class Outputs:
+        data = Output("Data", Table)
+
+    def __init__(self):
+        super().__init__()
+
+        # Interface utilisateur
+        self.access_key_input = QLineEdit(self)
+        self.access_key_input.setPlaceholderText("AWS Access Key")
+        self.access_key_input.setText(self.access_key)
+        self.access_key_input.editingFinished.connect(self.update_settings)
+        gui.widgetBox(self.controlArea, orientation='vertical').layout().addWidget(self.access_key_input)
+
+        self.secret_key_input = QLineEdit(self)
+        self.secret_key_input.setPlaceholderText("AWS Secret Key")
+        self.secret_key_input.setText(self.secret_key)
+        self.secret_key_input.editingFinished.connect(self.update_settings)
+        gui.widgetBox(self.controlArea, orientation='vertical').layout().addWidget(self.secret_key_input)
+
+        self.bucket_input = QLineEdit(self)
+        self.bucket_input.setPlaceholderText("Bucket S3")
+        self.bucket_input.setText(self.bucket_name)
+        self.bucket_input.editingFinished.connect(self.update_settings)
+        gui.widgetBox(self.controlArea, orientation='vertical').layout().addWidget(self.bucket_input)
+
+        self.extensions_input = QLineEdit(self)
+        self.extensions_input.setPlaceholderText("Extensions files from S3")
+        self.extensions_input.setText(",".join(self.allowed_extensions))
+        self.extensions_input.editingFinished.connect(self.update_settings)
+        gui.widgetBox(self.controlArea, orientation='vertical').layout().addWidget(self.extensions_input)
+
+        # self.folder_input = QLineEdit(self)
+        # self.folder_input.setPlaceholderText("Folder files from S3")
+        # self.folder_input.setText(self.folder_name)
+        # self.folder_input.editingFinished.connect(self.update_settings)
+        # gui.widgetBox(self.controlArea, orientation='vertical').layout().addWidget(self.extensions_input)
+
+        gui.button(self.controlArea, self, "Lister les fichiers", callback=self.list_files)
+
+    def update_settings(self):
+        self.access_key = self.access_key_input.text()
+        self.secret_key = self.secret_key_input.text()
+        self.bucket_name = self.bucket_input.text()
+        self.allowed_extensions = self.extensions_input.text()
+        self.folder_name = self.folder_input.text()
+
+    def list_files(self):
+        self.error(None)
+        try:
+            session = boto3.Session(
+                aws_access_key_id=self.access_key,
+                aws_secret_access_key=self.secret_key
+            )
+            s3 = session.client("s3")
+            response = s3.list_objects_v2(Bucket=self.bucket_name, Prefix = self.folder_name)
+            if self.allowed_extensions == "*":
+                self.allowed_extensions = "*"
+            elif isinstance(self.allowed_extensions, str):
+                self.allowed_extensions = tuple(ext.strip().lower() for ext in self.allowed_extensions.split(","))
+
+            if "Contents" in response:
+                data = []
+                meta = []
+                for obj in response["Contents"]:
+                    if self.allowed_extensions == "*" or self.allowed_extensions == "" or obj["Key"].lower().endswith(self.allowed_extensions):
+                        data.append([str(obj["Size"])])
+                        meta.append([obj["Key"],obj["LastModified"].strftime("%Y-%m-%d %H:%M:%S")])
+
+            domain = Domain(
+                [ContinuousVariable("Taille (bytes)")], metas=[StringVariable("Nom"), StringVariable("Date")]
+            )
+            orange_table = Table.from_numpy(domain, data, metas=meta)
+            self.Outputs.data.send(orange_table)
+
+        except Exception as e:
+            self.error(str(e))
+
+if __name__ == "__main__":
+    import sys
+
+    app = QApplication(sys.argv)
+    window = OWS3FileLister()
+    window.show()
+    sys.exit(app.exec_())

IO4IT-0.0.0.1/orangecontrib/IOIT/widgets/OWwordpdf2docx.py

@@ -0,0 +1,126 @@
+import os
+import sys
+
+import Orange.data
+from AnyQt.QtWidgets import QApplication
+from Orange.widgets import widget
+from Orange.widgets.utils.signals import Input, Output
+
+
+if "site-packages/Orange/widgets" in os.path.dirname(os.path.abspath(__file__)).replace("\\", "/"):
+    from Orange.widgets.orangecontrib.AAIT.utils import thread_management
+    from Orange.widgets.orangecontrib.AAIT.utils.import_uic import uic
+    from Orange.widgets.orangecontrib.AAIT.utils.initialize_from_ini import apply_modification_from_python_file
+    from Orange.widgets.orangecontrib.IOIT.ocr_function import word_converter
+else:
+    from orangecontrib.AAIT.utils import thread_management
+    from orangecontrib.AAIT.utils.import_uic import uic
+    from orangecontrib.AAIT.utils.initialize_from_ini import apply_modification_from_python_file
+    from orangecontrib.IOIT.ocr_function import word_converter
+
+@apply_modification_from_python_file(filepath_original_widget=__file__)
+class OWCreateEmbeddings(widget.OWWidget):
+    name = "WordPdf2Docx"
+    description = "Convert pdf from a directory to docx using word"
+    icon = "icons/wordpdf2docx.png"
+    if "site-packages/Orange/widgets" in os.path.dirname(os.path.abspath(__file__)).replace("\\", "/"):
+        icon = "icons_dev/wordpdf2docx.png"
+    gui = os.path.join(os.path.dirname(os.path.abspath(__file__)), "designer/wordpdf2docx.ui")
+    want_control_area = False
+    priority = 3000
+
+    class Inputs:
+        data = Input("Data", Orange.data.Table)
+
+    class Outputs:
+        data = Output("Data", Orange.data.Table)
+
+    @Inputs.data
+    def set_data(self, in_data):
+        self.data = in_data
+        if self.autorun:
+            self.run()
+
+
+    def __init__(self):
+        super().__init__()
+        # Qt Management
+        self.setFixedWidth(470)
+        self.setFixedHeight(300)
+        uic.loadUi(self.gui, self)
+
+        # Data Management
+        self.data = None
+        self.thread = None
+        self.autorun = True
+        self.result = None
+        self.post_initialized()
+
+    def run(self):
+        # if thread is running quit
+        if self.thread is not None:
+            self.thread.safe_quit()
+
+        if self.data is None:
+            return
+
+
+        # Verification of in_data
+        self.error("")
+        try:
+            self.data.domain["input_dir"]
+        except KeyError:
+            self.error('You need a "input_dir" column in input data')
+            return
+
+        if type(self.data.domain["input_dir"]).__name__ != 'StringVariable':
+            self.error('"input_dir" column needs to be a Text')
+            return
+        try:
+            self.data.domain["output_dir"]
+        except KeyError:
+            self.error('You need a "output_dir" column in input data')
+            return
+
+        if type(self.data.domain["output_dir"]).__name__ != 'StringVariable':
+            self.error('"output_dir" column needs to be a Text')
+            return
+        input_dir = self.data.get_column("input_dir")
+        output_dir = self.data.get_column("output_dir")
+
+        # Start progress bar
+        self.progressBarInit()
+        # Connect and start thread : main function, progress, result and finish
+        # --> progress is used in the main function to track progress (with a callback)
+        # --> result is used to collect the result from main function
+        # --> finish is just an empty signal to indicate that the thread is finished
+        self.thread = thread_management.Thread(word_converter.convert_pdf_structure, input_dir, output_dir)
+        self.thread.progress.connect(self.handle_progress)
+        self.thread.result.connect(self.handle_result)
+        self.thread.finish.connect(self.handle_finish)
+        self.thread.start()
+
+    def handle_progress(self, value: float) -> None:
+        self.progressBarSet(value)
+
+    def handle_result(self, result):
+        try:
+            self.result = result
+            self.Outputs.data.send(self.data)
+        except Exception as e:
+            print("An error occurred when sending out_data:", e)
+            self.Outputs.data.send(None)
+            return
+
+    def handle_finish(self):
+        print("conversion finished")
+        self.progressBarFinished()
+
+    def post_initialized(self):
+        pass
+
+if __name__ == "__main__":
+    app = QApplication(sys.argv)
+    my_widget = OWCreateEmbeddings()
+    my_widget.show()
+    app.exec_()

IO4IT-0.0.0.1/orangecontrib/IOIT/widgets/__init__.py

@@ -0,0 +1,19 @@
+"""
+Widgets from Development workflow category
+
+"""
+# ID = "orangecontrib.AAIT"
+
+NAME = "Advanced Artificial Intelligence Tools"
+
+# Category icon show in the menu
+ICON = "icons/category.svg"
+
+BACKGROUND = "light-green"
+
+DESCRIPTION = ("Advanced Artificial Intelligence Tools is a package meant to develop and enable advanced AI "
+               "functionalities in Orange Data Mining.")
+
+
+
+# PRIORITY = 6

IO4IT-0.0.0.1/orangecontrib/IOIT/widgets/designer/wordpdf2docx.ui

@@ -0,0 +1,41 @@
+<?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>300</height>
+   </rect>
+  </property>
+  <property name="windowTitle">
+   <string>Form</string>
+  </property>
+  <widget class="QLabel" name="Description">
+   <property name="geometry">
+    <rect>
+     <x>20</x>
+     <y>20</y>
+     <width>431</width>
+     <height>261</height>
+    </rect>
+   </property>
+   <property name="text">
+    <string>convert pdf file to docx from input_dir to output_dir</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>
+ <resources/>
+ <connections/>
+</ui>

IO4IT-0.0.0.1/orangecontrib/IOIT/widgets/icons/category.svg

@@ -0,0 +1,50 @@
+<?xml version="1.0" ?>
+
+<svg width="800px" height="800px" viewBox="0 0 48 48" data-name="Layer 1" id="Layer_1" xmlns="http://www.w3.org/2000/svg">
+
+
+      .cls-1, .cls-2 {
+        fill: none;
+        stroke: #353535;
+        stroke-linecap: round;
+        stroke-width: 2px;
+      }
+
+      .cls-1 {
+        stroke-linejoin: round;
+      }
+
+      .cls-2 {
+        stroke-miterlimit: 10;
+      }
+
+      .cls-3 {
+        fill: #353535;
+      }
+    </style>
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+

IO4IT-0.0.0.1/orangecontrib/IOIT/widgets/ocr_function/word_converter.py

@@ -0,0 +1,81 @@
+import os
+import win32com.client
+from pathlib import Path
+
+#docx_path = pdf_path.replace(".pdf", ".docx")
+def convert_pdf_structure(input_dir: str, output_dir: str, progress_callback=None):
+    """
+    return a string with log in  case of error
+    Recursively lists all .pdf and .PDF files in the input directory,
+    replicates the folder structure in the output directory, and
+    creates empty .docx files with the same names.
+
+    Parameters:
+    input_dir (str): Path to the input directory containing PDF files.
+    output_dir (str): Path to the output directory where DOCX files will be created.
+    """
+    error_log=""
+    if os.name != 'nt':
+        error_log="version developped for windows computer "
+        return error_log
+
+    nbre_file = 0
+    for i, data in enumerate(input_dir):
+        input_path = Path(str(input_dir[i]))
+        for pdf_file in input_path.rglob("*.pdf"):
+            nbre_file += 1
+
+    k = 1
+    for i, data in enumerate(input_dir):
+        input_path = Path(str(input_dir[i]))
+        output_path = Path(str(output_dir[i]))
+
+        if not input_path.exists() or not input_path.is_dir():
+            print(f"Error: The input directory '{input_dir}' does not exist or is not a directory.")
+            return f"Error: The input directory '{input_dir}' does not exist or is not a directory. "
+
+        for pdf_file in input_path.rglob("*.pdf"):  # Recursively search for .pdf and .PDF files
+            relative_path = pdf_file.relative_to(input_path)  # Get relative path from input root
+            new_file_path = output_path / relative_path.with_suffix(".docx")  # Change extension to .docx
+
+            # Create necessary directories in the output folder
+            new_file_path.parent.mkdir(parents=True, exist_ok=True)
+            if 0!= convert_pdf_to_docx(str(pdf_file),str(new_file_path)):
+                if error_log!="":
+                    error_log+="\n"
+                error_log+="error -> "+pdf_file
+            if progress_callback is not None:
+                progress_value = float(100 * (k) / nbre_file)
+                k += 1
+                progress_callback(progress_value)
+    return error_log
+
+def convert_pdf_to_docx(pdf_path,docx_path):
+    # return 0 ok return 1 ok
+    to_return = 1
+    if not os.path.exists(pdf_path):
+        print(f"File {pdf_path} doesn t exist.")
+        return to_return
+    try:
+        # Lancer Word
+        word = win32com.client.Dispatch("Word.Application")
+        word.Visible = True  # Mettre à True pour voir Word en action
+        word.DisplayAlerts = 0  # Désactiver les alertes et fenêtres
+        print(f"Conversion  {pdf_path} to {docx_path}...")
+
+        # Ouvrir le PDF avec les bons paramètres
+        doc = word.Documents.Open(pdf_path, ReadOnly=True, ConfirmConversions=False)
+
+        # Sauvegarder en DOCX
+        doc.SaveAs(docx_path, FileFormat=16)  # 16 = wdFormatDocumentDefault
+        doc.Close(False)
+
+        print(f"Conversion ok : {docx_path}")
+        to_return=0
+    except Exception as e:
+        print(f"Error : {e}")
+
+
+    finally:
+        word.Quit()
+        return to_return

IO4IT-0.0.0.1/setup.cfg

@@ -0,0 +1,4 @@
+[egg_info]
+tag_build = 
+tag_date = 0
+

IO4IT-0.0.0.1/setup.py

@@ -0,0 +1,58 @@
+from setuptools import setup, find_packages
+
+# Nom du package PyPI ('pip install NAME')
+NAME = "IO4IT"
+
+# Version du package PyPI
+VERSION = "0.0.0.1"  # la version doit être supérieure à la précédente sinon la publication sera refusée
+
+# Facultatif / Adaptable à souhait
+AUTHOR = ""
+AUTHOR_EMAIL = ""
+URL = ""
+DESCRIPTION = ""
+LICENSE = ""
+
+# 'orange3 add-on' permet de rendre l'addon téléchargeable via l'interface addons d'Orange
+KEYWORDS = ["orange3 add-on",]
+
+# Tous les packages python existants dans le projet (avec un __ini__.py)
+PACKAGES = find_packages()
+PACKAGES = [pack for pack in PACKAGES if "orangecontrib" in pack and "IOIT" in pack]
+print(PACKAGES)
+
+
+
+# Fichiers additionnels aux fichiers .py (comme les icons ou des .ows)
+PACKAGE_DATA = {
+    "orangecontrib.IOIT.widgets": ["icons/*", "designer/*"],
+}
+# /!\ les noms de fichier 'orangecontrib.hkh_bot.widgets' doivent correspondre à l'arborescence
+
+# Dépendances
+INSTALL_REQUIRES = ["boto3", "fastapi"]
+
+# Spécifie le dossier contenant les widgets et le nom de section qu'aura l'addon sur Orange
+ENTRY_POINTS = {
+    "orange.widgets": (
+        "Advanced Artificial Intelligence Tools = orangecontrib.AAIT.widgets",
+    )
+}
+# /!\ les noms de fichier 'orangecontrib.hkh_bot.widgets' doivent correspondre à l'arborescence
+
+NAMESPACE_PACKAGES = ["orangecontrib"]
+
+setup(name=NAME,
+      version=VERSION,
+      author=AUTHOR,
+      author_email=AUTHOR_EMAIL,
+      url=URL,
+      description=DESCRIPTION,
+      license=LICENSE,
+      keywords=KEYWORDS,
+      packages=PACKAGES,
+      package_data=PACKAGE_DATA,
+      install_requires=INSTALL_REQUIRES,
+      entry_points=ENTRY_POINTS,
+      namespace_packages=NAMESPACE_PACKAGES,
+      )