IO4IT 2.1.0__tar.gz → 2.1.0.4__tar.gz

io4it-2.1.0.4/PKG-INFO

@@ -0,0 +1,33 @@
+Metadata-Version: 2.1
+Name: io4it
+Version: 2.1.0.4
+Summary: UNKNOWN
+Home-page: 
+Author: 
+Author-email: 
+License: UNKNOWN
+Keywords: orange3 add-on
+Platform: UNKNOWN
+Requires-Dist: pylatexenc
+Requires-Dist: docopt
+Requires-Dist: boto3
+Requires-Dist: opencv-python-headless==4.6.0.66
+Requires-Dist: docling==2.30.0
+Requires-Dist: docling-core==2.26.3
+Requires-Dist: speechbrain
+Requires-Dist: whisper
+Requires-Dist: whisper-openai
+Requires-Dist: pyannote.audio
+Requires-Dist: pyannote-core
+Requires-Dist: pypandoc
+Requires-Dist: pypandoc-binary
+Requires-Dist: scikit-learn
+Requires-Dist: openai
+Requires-Dist: pip-system-certs==5.0
+Requires-Dist: docx2pdf
+Requires-Dist: doc2docx
+Requires-Dist: msal
+Requires-Dist: exchangelib
+Requires-Dist: CATEGORIT
+
+UNKNOWN

io4it-2.1.0.4/io4it.egg-info/PKG-INFO

@@ -0,0 +1,33 @@
+Metadata-Version: 2.1
+Name: io4it
+Version: 2.1.0.4
+Summary: UNKNOWN
+Home-page: 
+Author: 
+Author-email: 
+License: UNKNOWN
+Keywords: orange3 add-on
+Platform: UNKNOWN
+Requires-Dist: pylatexenc
+Requires-Dist: docopt
+Requires-Dist: boto3
+Requires-Dist: opencv-python-headless==4.6.0.66
+Requires-Dist: docling==2.30.0
+Requires-Dist: docling-core==2.26.3
+Requires-Dist: speechbrain
+Requires-Dist: whisper
+Requires-Dist: whisper-openai
+Requires-Dist: pyannote.audio
+Requires-Dist: pyannote-core
+Requires-Dist: pypandoc
+Requires-Dist: pypandoc-binary
+Requires-Dist: scikit-learn
+Requires-Dist: openai
+Requires-Dist: pip-system-certs==5.0
+Requires-Dist: docx2pdf
+Requires-Dist: doc2docx
+Requires-Dist: msal
+Requires-Dist: exchangelib
+Requires-Dist: CATEGORIT
+
+UNKNOWN

{io4it-2.1.0 → io4it-2.1.0.4}/io4it.egg-info/SOURCES.txt

@@ -13,6 +13,8 @@ orangecontrib/IO4IT/ocr_function/word_converter.py
 orangecontrib/IO4IT/utils/__init__.py
 orangecontrib/IO4IT/utils/mail.py
 orangecontrib/IO4IT/utils/offuscation_basique.py
+orangecontrib/IO4IT/utils/pool_exec_utils.py
+orangecontrib/IO4IT/utils/utils_md.py
 orangecontrib/IO4IT/widgets/OWChatGpt.py
 orangecontrib/IO4IT/widgets/OWDeep_Search.py
 orangecontrib/IO4IT/widgets/OWExportMarkdown.py

{io4it-2.1.0 → io4it-2.1.0.4}/orangecontrib/IO4IT/utils/mail.py

@@ -4,19 +4,21 @@ import email
 from email.header import decode_header
 import time
 import os
-from bs4 import BeautifulSoup
 from email import policy
 from email.message import EmailMessage
 import mimetypes
 
-from exchangelib import Credentials, Account, DELEGATE, Configuration, HTMLBody, Message, Mailbox
+from exchangelib import Credentials, HTMLBody, Message, Mailbox
 from exchangelib.protocol import BaseProtocol, NoVerifyHTTPAdapter
 from exchangelib import OAuth2Credentials, Identity, Configuration, Account, DELEGATE
-from oauthlib.oauth2 import OAuth2Token
+
 
 from bs4 import BeautifulSoup
-BaseProtocol.HTTP_ADAPTER_CLS=NoVerifyHTTPAdapter
-import json
+
+
+##### please uncomment this line to ignore certificate checking (owa)
+#BaseProtocol.HTTP_ADAPTER_CLS=NoVerifyHTTPAdapter
+
 from msal import ConfidentialClientApplication
 from oauthlib.oauth2 import OAuth2Token
 
@@ -185,6 +187,8 @@ def check_new_emails(offusc_conf_agent,type_co, list_agent_email=[]):
                                 with open(folder + "/" + "mail.ok", "w") as f:
                                     f.close()
 
+
+
             imap.logout()
         except Exception as e:
             print(f"Erreur lors de la vérification des mails : {e}")
@@ -199,7 +203,7 @@ def check_new_emails(offusc_conf_agent,type_co, list_agent_email=[]):
             )
 
             account = Account(
-                primary_smtp_address=mail,
+                primary_smtp_address=alias,
                 credentials=credentials,
                 config=config,
                 autodiscover=False,
@@ -301,6 +305,7 @@ def check_new_emails(offusc_conf_agent,type_co, list_agent_email=[]):
                         f.write("\n".join(output_lines))
                     with open(os.path.join(folder, "mail.ok"), "w") as f:
                         pass
+
         except Exception as e:
             print(f"Erreur lors du traitement du mail : {e}")
     elif type_co == "MICROSOFT_EXCHANGE_OAUTH2":
@@ -427,6 +432,7 @@ def check_new_emails(offusc_conf_agent,type_co, list_agent_email=[]):
                     with open(os.path.join(folder, "mail.ok"), "w") as f:
                         pass
 
+
         except Exception as e:
             print(f"Erreur OAuth2 lors du traitement du mail : {e}")
 
@@ -662,13 +668,27 @@ def check_send_new_emails(offusc_conf_agent,type_co):
     else:
         print("type de co non valide")
 
+def list_conf_files(type_co):
+    conf_files = []
+    if type_co is None or type_co == "":
+        return conf_files
+    dossier = MetManagement.get_secret_content_dir()
+    dossier = dossier + type_co
+    files = os.listdir(dossier)
+    for file in files:
+        if file.lower().endswith(".json"):
+            conf_files.append(file)
+    return conf_files
+
+
+
 if __name__ == "__main__":
-    offusc_conf_agent=""
-    #type_co="IMAP4_SSL"
-    type_co="MICROSOFT_EXCHANGE_OAUTH2"
+    type_co="IMAP4_SSL"
     list_agent_email = []
+    offusc_conf_agents = list_conf_files(type_co)
     while True:
-        check_new_emails(offusc_conf_agent,type_co, list_agent_email)
-        time.sleep(1)
-        check_send_new_emails(offusc_conf_agent,type_co)
-        time.sleep(1)
+        for offusc_conf_agent in offusc_conf_agents:
+            check_new_emails(offusc_conf_agent,type_co, list_agent_email)
+            time.sleep(1)
+            check_send_new_emails(offusc_conf_agent,type_co)
+            time.sleep(1)

{io4it-2.1.0 → io4it-2.1.0.4}/orangecontrib/IO4IT/utils/offuscation_basique.py

@@ -2,6 +2,9 @@ import os
 import json
 import hashlib
 import getpass
+from cryptography.fernet import Fernet
+import base64
+
 
 if "site-packages/Orange/widgets" in os.path.dirname(os.path.abspath(__file__)).replace("\\", "/"):
     from Orange.widgets.orangecontrib.AAIT.utils import MetManagement
@@ -47,12 +50,26 @@ def get_keys_dir(type_key: str = "MICROSOFT_EXCHANGE_OAUTH2") -> str:
         if os.path.basename(base) != "keys":
             base = os.path.join(base, "keys")
         dossier = os.path.normpath(os.path.join(base, type_key))
-        print("dossier", dossier)
         os.makedirs(dossier, exist_ok=True)
         return dossier
     except Exception as e:
         raise RuntimeError(f"Erreur création/récupération dossier : {e}")
 
+def get_fernet_key() -> bytes:
+    """
+    Dérive une clé Fernet (32 octets base64) à partir du nom d'utilisateur local.
+    """
+    username = getpass.getuser().encode("utf-8")
+    digest = hashlib.sha256(username).digest()
+    return base64.urlsafe_b64encode(digest[:32])  # 32 bytes en base64
+
+def encrypt_secure(data: str) -> str:
+    fernet = Fernet(get_fernet_key())
+    return fernet.encrypt(data.encode("utf-8")).decode("utf-8")
+
+def decrypt_secure(data: str) -> str:
+    fernet = Fernet(get_fernet_key())
+    return fernet.decrypt(data.encode("utf-8")).decode("utf-8")
 
 
 # GSTION IMAP
@@ -160,6 +177,83 @@ def enregistrer_config_owa(mail,alias,server,username,password,interval):
         print(f"❌ Erreur lors de l'enregistrement : {e}")
         return 1
 
+def enregistrer_config_owa_secure(mail, alias, server, username, password, interval):
+    try:
+        dossier = get_keys_dir("MICROSOFT_EXCHANGE_OWA_SECURE")
+        os.makedirs(dossier, exist_ok=True)
+
+        mdp_chiffre = encrypt_secure(password)
+
+        nom_fichier = os.path.join(dossier, f"{alias.replace('@', '_at_')}.json")
+
+        contenu = {
+            "mail": mail,
+            "alias": alias,
+            "server": server,
+            "username": username,
+            "password_encrypted": mdp_chiffre,
+            "interval_second": interval
+        }
+
+        with open(nom_fichier, "w", encoding="utf-8") as f:
+            json.dump(contenu, f, indent=4)
+
+        print(f"✅ Fichier OWA sécurisé enregistré : {nom_fichier}")
+        return 0
+
+    except Exception as e:
+        print(f"❌ Erreur d'enregistrement sécurisé OWA : {e}")
+        return 1
+
+def enregistrer_config_cli_owa_secure():
+    print("\n🔐 Écriture fichier OWA avec chiffrement sécurisé :")
+    mail = input("📧 Mail (nom@domain.com) : ").strip()
+    alias = mail #input("📛 Alias (visible) : ").strip()
+    server = input("🌐 Server : ").strip()
+    username = input("👤 Username (domain\\user) : ").strip()
+    password = input("🔑 Password (non masqué) : ").strip()
+    interval = int(input("⏱️ Interval (secondes) : ").strip())
+    if alias == "''" or alias == "\"\"" or alias == "":
+        alias = mail
+
+    if 0 != enregistrer_config_owa_secure(mail, alias, server, username, password, interval):
+        print("❌ Erreur lors de enregistrer_config_owa_secure !")
+
+def lecture_config_cli_owa_secure():
+    fichier = input("📄 Nom du fichier JSON (sans chemin) : ").strip()
+    cfg = lire_config_owa_secure(fichier)
+
+    if cfg is None:
+        print("❌ Erreur")
+        return
+
+    print(f"\n📧 mail        : {cfg[0]}")
+    print(f"📛 alias       : {cfg[1]}")
+    print(f"🌐 server      : {cfg[2]}")
+    print(f"👤 username    : {cfg[3]}")
+    print(f"🔑 password    : {cfg[4]}")
+    print(f"⏱️ intervalle  : {cfg[5]}s")
+
+def lire_config_owa_secure(chemin_fichier):
+    try:
+        chemin_fichier = os.path.join(get_keys_dir("MICROSOFT_EXCHANGE_OWA_SECURE"), chemin_fichier)
+        with open(chemin_fichier, "r", encoding="utf-8") as f:
+            contenu = json.load(f)
+
+        mdp_dechiffre = decrypt_secure(contenu["password_encrypted"])
+
+        return [
+            contenu["mail"],
+            contenu["alias"],
+            contenu["server"],
+            contenu["username"],
+            mdp_dechiffre,
+            int(contenu["interval_second"])
+        ]
+
+    except Exception as e:
+        print(f"❌ Erreur de lecture sécurisée OWA : {e}")
+        return None
 
 
 # Fonction pour lire le fichier de configuration et déchiffrer le mot de passe
@@ -357,21 +451,28 @@ def lire_config_api(service_name):
 def enregistrer_config_cli_api():
     print("\n📝 Écriture d’une clé API :")
     service = input("🔖 Nom du service : ").strip()
-    api_key = getpass.getpass("🔑 Clé API         : ").strip()
+    api_key = input("🔑 Clé API         : ").strip()
     desc = input("✏️  Description      : ").strip()
     if 0 != enregistrer_config_api(service, api_key, desc):
         print("erreur!")
 
-def lire_config_cli_api():
-    service = input("🔖 Nom du service : ").strip()
-    cfg = lire_config_api(service)
-    if cfg is None:
-        print("erreur")
-        return
-    print(f"\n📄 service     : {cfg['service']}")
-    print(f"🔑 clé API     : {cfg['api_key']}")
-    if cfg['description']:
-        print(f"📝 description : {cfg['description']}")
+def lire_config_cli_api(service=""):
+    if service == "":
+        service = input("🔖 Nom du service : ").strip()
+    try:
+        cfg = lire_config_api(service)
+        if cfg is None:
+            print("erreur")
+            return
+        print(f"\n📄 service     : {cfg['service']}")
+        print(f"🔑 clé API     : {cfg['api_key']}")
+        if cfg['description']:
+            print(f"📝 description : {cfg['description']}")
+        return cfg['api_key']
+    except Exception as e:
+        print(f"❌ Erreur lors de la lecture : {e}")
+        return None
+
 
 
 # Gestion d’éléments de nxp (DOSSIER_NODE_ID, SERVEUR, USERNAME, PASSWORD)  (HARD dossier aait_store/keys)
@@ -458,8 +559,6 @@ def lire_config_cli_nxp():
         print(f"📝 description     : {cfg['description']}")
 
 
-
-
 if __name__ == "__main__":
     print("1) ecrire fichier IMAP4_SSL")
     print("2) dechiffer fichier IMAP4_SSL")
@@ -471,7 +570,10 @@ if __name__ == "__main__":
     print("8) Déchiffrer fichier Microsoft Exchange (OWA)")
     print("9) Écrire fichier Microsoft Exchange (OAuth2)")
     print("10) Déchiffrer fichier Microsoft Exchange (OAuth2)")
-    choix = input("👉 Que faire ? [1-10] : ").strip()
+    print("11) Écrire fichier Microsoft Exchange (OWA) [SECURE]")
+    print("12) Déchiffrer fichier Microsoft Exchange (OWA) [SECURE]")
+
+    choix = input("👉 Que faire ? [1-12] : ").strip()
 
     if choix == "1":
         enregistrer_config_cli_imap4_ssl()
@@ -492,7 +594,12 @@ if __name__ == "__main__":
     elif choix=="9":
         enregistrer_config_cli_oauth2()
     elif choix=="10":
-        lecture_config_cli_oauth2()
+        lire_config_cli_oauth2()
+    elif choix == "11":
+        enregistrer_config_cli_owa_secure()
+    elif choix == "12":
+        lecture_config_cli_owa_secure()
+
 
     else:
         print("❌ Choix invalide. Réessayez.\n")

io4it-2.1.0.4/orangecontrib/IO4IT/utils/pool_exec_utils.py

@@ -0,0 +1,33 @@
+# pool_exec_utils.py
+import os
+from concurrent.futures import ProcessPoolExecutor
+from typing import Optional, Tuple
+
+def available_cpus(default: int = 4) -> int:
+    c = os.cpu_count()
+    return c if (c and c > 0) else default
+
+def cpu_label_text() -> str:
+    return f"🖥️ CPU dispo : {available_cpus()}"
+
+def shutdown_executor(executor: Optional[ProcessPoolExecutor]) -> None:
+    if executor is not None:
+        try:
+            executor.shutdown(wait=False, cancel_futures=True)
+        except TypeError:
+            executor.shutdown(wait=False)
+        except Exception:
+            pass
+
+def create_or_update_executor(
+    executor: Optional[ProcessPoolExecutor],
+    current_workers: Optional[int],
+    new_workers: int,
+) -> Tuple[ProcessPoolExecutor, int, str, bool]:
+    changed = (executor is None) or (current_workers != int(new_workers))
+    if changed:
+        shutdown_executor(executor)
+        executor = ProcessPoolExecutor(max_workers=int(new_workers))
+        return executor, int(new_workers), f"Executor prêt ({int(new_workers)} workers).", True
+    else:
+        return executor, int(current_workers), f"Executor inchangé ({int(new_workers)} workers).", False

io4it-2.1.0.4/orangecontrib/IO4IT/utils/utils_md.py

@@ -0,0 +1,168 @@
+# utils_md.py
+import os, re, hashlib, secrets, urllib.parse, logging, time, datetime, numpy as np
+from pathlib import Path
+import fitz  # PyMuPDF
+
+try:
+    import easyocr
+except Exception:
+    easyocr = None
+
+MAX_STEM_LEN = 40
+IMAGE_RESOLUTION_SCALE = 2.0
+
+_log = logging.getLogger(__name__)
+logging.getLogger("fitz").setLevel(logging.ERROR)
+logging.getLogger("PIL").setLevel(logging.WARNING)
+
+def slugify(text: str) -> str:
+    text = text.strip().lower()
+    text = re.sub(r"[^\w\-]+", "-", text)
+    text = re.sub(r"-{2,}", "-", text)
+    return text.strip("-") or "file"
+
+def short_stem(original_stem: str) -> str:
+    base = slugify(original_stem)[:MAX_STEM_LEN]
+    h = secrets.token_hex(6)
+    return f"{base}-{h}"
+
+def truncate_path(path, max_length=240):
+    p = Path(path)
+    s = str(p)
+    if len(s) <= max_length:
+        return p
+    h = hashlib.md5(s.encode()).hexdigest()
+    new_name = (p.stem[:50] + "_" + h + p.suffix)
+    return p.parent / new_name
+
+def try_read_text(path: Path) -> str:
+    for enc in ("utf-8", "utf-8-sig", "latin-1"):
+        try:
+            return urllib.parse.unquote(path.read_text(encoding=enc, errors="ignore"))
+        except Exception:
+            continue
+    return ""
+
+def is_pdf_text_based(fpath: Path) -> bool:
+    try:
+        with fitz.open(fpath) as doc:
+            for page in doc:
+                if page.get_text().strip():
+                    return True
+            return False
+    except Exception:
+        return False
+
+def ocr_fallback(pdf_path: Path, langs=('fr','en')) -> str:
+    if easyocr is None:
+        raise RuntimeError("easyocr introuvable. Installez easyocr pour l'OCR.")
+    reader = easyocr.Reader(list(langs))
+    _ = reader.readtext(np.zeros((10, 10, 3), dtype=np.uint8), detail=0)  # warm-up
+    text = []
+    with fitz.open(pdf_path) as doc:
+        for page in doc:
+            pix = page.get_pixmap(dpi=200)
+            img = np.frombuffer(pix.samples, dtype=np.uint8).reshape((pix.height, pix.width, pix.n))
+            result = reader.readtext(img)
+            text.extend([r[1] for r in result])
+    return "\n".join(text)
+
+def strip_image_markdown(md_text: str) -> str:
+    md_text = re.sub(r'!\[[^\]]*\]\([^\)]+\)', '', md_text)
+    md_text = re.sub(r'<img\b[^>]*>', '', md_text, flags=re.IGNORECASE)
+    md_text = re.sub(r'\n{3,}', '\n\n', md_text)
+    return md_text.strip()
+
+# --- COM helpers (Windows + Office)
+def _make_word_invisible(word):
+    try:
+        word.Visible = False
+        word.DisplayAlerts = 0
+        word.ScreenUpdating = False
+    except Exception:
+        pass
+
+def _make_powerpoint_invisible(ppt):
+    try:
+        ppt.Visible = 0  # 0=Hidden
+    except Exception:
+        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
+    dst = out_dir / (src.stem + ".docx")
+    if dst.exists() and dst.stat().st_size > 0:
+        return dst
+    word = win32com.client.Dispatch("Word.Application")
+    _make_word_invisible(word)
+    try:
+        doc = word.Documents.Open(str(src), ReadOnly=True)
+        doc.SaveAs(str(dst), FileFormat=16)  # DOCX
+        doc.Close(False)
+        if not dst.exists() or dst.stat().st_size == 0:
+            raise RuntimeError("DOCX non généré")
+        return dst
+    finally:
+        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
+    dst = out_dir / (src.stem + ".pptx")
+    if dst.exists() and dst.stat().st_size > 0:
+        return dst
+    ppt = win32com.client.Dispatch("PowerPoint.Application")
+    _make_powerpoint_invisible(ppt)
+    try:
+        pres = ppt.Presentations.Open(str(src), WithWindow=False)
+        pres.SaveAs(str(dst), 24)  # ppSaveAsOpenXMLPresentation
+        pres.Close()
+        if not dst.exists() or dst.stat().st_size == 0:
+            raise RuntimeError("PPTX non généré")
+        return dst
+    finally:
+        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:
+        return pdf_path
+    word = win32com.client.Dispatch("Word.Application")
+    _make_word_invisible(word)
+    try:
+        doc = word.Documents.Open(str(src_docx), ReadOnly=True)
+        try:
+            doc.ExportAsFixedFormat(OutputFileName=str(pdf_path), ExportFormat=17)
+        except Exception:
+            doc.SaveAs(str(pdf_path), FileFormat=17)
+        doc.Close(False)
+        if not pdf_path.exists() or pdf_path.stat().st_size == 0:
+            raise RuntimeError("PDF non généré depuis DOCX")
+        return pdf_path
+    finally:
+        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:
+        return pdf_path
+    ppt = win32com.client.Dispatch("PowerPoint.Application")
+    _make_powerpoint_invisible(ppt)
+    try:
+        pres = ppt.Presentations.Open(str(src_pptx), WithWindow=False)
+        pres.SaveAs(str(pdf_path), 32)  # PDF
+        pres.Close()
+        if not pdf_path.exists() or pdf_path.stat().st_size == 0:
+            raise RuntimeError("PDF non généré depuis PPTX")
+        return pdf_path
+    finally:
+        ppt.Quit()

{io4it-2.1.0 → io4it-2.1.0.4}/orangecontrib/IO4IT/widgets/OWExportMarkdown.py

@@ -165,7 +165,7 @@ class OWExportMarkdown(widget.OWWidget):
 
             try:
                 # DOCX
-                pypandoc.convert_file(tmp_md, to="docx", outputfile=docx_out)
+                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",
@@ -173,7 +173,7 @@ class OWExportMarkdown(widget.OWWidget):
                 )
 
                 # PPTX
-                pypandoc.convert_file(tmp_md, to="pptx", outputfile=pptx_out)
+                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",