fixed error in compose.yml

fixed regular deletions of files
made to kill fip entrys when 1d passes
2025-11-04 21:24:38 +01:00 · 2025-11-04 19:16:59 +01:00 · 2025-11-04 19:04:41 +01:00 · 2025-11-04 17:54:24 +01:00 · 2025-11-04 17:17:13 +01:00 · 2025-11-04 17:13:18 +01:00
28 changed files with 1358 additions and 305 deletions
--- a/.gitignore
+++ b/.gitignore
@@ -3,4 +3,9 @@ app/files/
 app/pwfile.json
 app/dest
 app.log
 init.log
 app/__pycache__/
 mariadb/*
 unizeug
 .mypy_cache
 .nvim
--- a/.nvim/rsync.toml_alt
+++ b/.nvim/rsync.toml_alt
@@ -1,2 +0,0 @@
 # remote_path="/srv/http/"
 # remote_path="dev@10.0.0.25:/var/www/html/"
--- a/31
+++ b/31
@@ -0,0 +1,31 @@
 FROM python:3.13-rc-alpine
 WORKDIR /usr/src/
 COPY requirements.txt /usr/src/requirements.txt
 COPY entrypoint.sh /usr/src/entrypoint.sh
 RUN apk add --no-cache \
    gcc \
    g++ \
    musl-dev \
    python3-dev \
    libffi-dev \
    openssl-dev \
    cargo \
    make \
    mariadb-connector-c-dev \
    jpeg-dev \
    zlib-dev \
    freetype-dev \
    lcms2-dev \
    openjpeg-dev \
    tiff-dev \
    tk-dev \
    tcl-dev \
    libwebp-dev \
    tesseract-ocr \
    tesseract-ocr-data-deu
 RUN python -m ensurepip --upgrade
 RUN pip install setuptools wheel
 RUN pip install -r requirements.txt
 WORKDIR /python
 CMD /bin/sh /usr/src/entrypoint.sh
 # ENTRYPOINT ["/usr/src/entrypoint.sh"]
--- a/app/pycache/main.cpython-312.pyc
+++ b/app/pycache/main.cpython-312.pyc
--- a/app/pycache/main.cpython-313.pyc
+++ b/app/pycache/main.cpython-313.pyc
--- a/app/favicon/apple-touch-icon.png
+++ b/app/favicon/apple-touch-icon.png
--- a/app/favicon/favicon-96x96.png
+++ b/app/favicon/favicon-96x96.png
--- a/app/favicon/favicon.ico
+++ b/app/favicon/favicon.ico
--- a/app/favicon/favicon.svg
+++ b/app/favicon/favicon.svg
--- a/app/favicon/generated_from.txt
+++ b/app/favicon/generated_from.txt
--- a/app/favicon/site.webmanifest
+++ b/app/favicon/site.webmanifest
--- a/app/favicon/web-app-manifest-192x192.png
+++ b/app/favicon/web-app-manifest-192x192.png
--- a/app/favicon/web-app-manifest-512x512.png
+++ b/app/favicon/web-app-manifest-512x512.png
--- a/app/graphics/greeting.pdf
+++ b/app/graphics/greeting.pdf
--- a/app/graphics/greeting.svg
+++ b/app/graphics/greeting.svg
--- a/app/index.html
+++ b/app/index.html
@@ -0,0 +1,164 @@
 <!doctype html>
 <html lang="de">
  <head>
    <title>Unizeug uploader</title>
    <link rel="stylesheet" href="static/style.css" />
    <script src="https://cdnjs.cloudflare.com/ajax/libs/pdf.js/2.4.456/pdf.min.js"></script>
    <script src="static/app.js" defer></script>
    <script src="static/autocomplete.js" defer></script>
    <script src="static/dynhide.js" defer></script>
    <script src="static/filedrop.js" defer></script>
    <link
      rel="icon"
      type="image/png"
      href="/favicon/favicon-96x96.png"
      sizes="96x96"
    />
    <link rel="icon" type="image/svg+xml" href="/favicon/favicon.svg" />
    <link rel="shortcut icon" href="/favicon/favicon.ico" />
    <link
      rel="apple-touch-icon"
      sizes="180x180"
      href="/favicon/apple-touch-icon.png"
    />
    <meta name="apple-mobile-web-app-title" content="Unizeug" />
    <link rel="manifest" href="/favicon/site.webmanifest" />
  </head>
  <body>
    <!-- The Modal -->
    <div id="loading" class="modal">
      <!-- Modal content -->
      <div class="loading-content">
        <!-- <span class="close">&times;</span> -->
        <div class="loader"></div>
        <p id="upload_status" class="upload_status_text">Uploading</p>
      </div>
    </div>
    <!-- main -->
    <div class="main">
      <div class="left" id="controldiv">
        <div id="fileupload">
          <form id="uploadform" enctype="multipart/form-data">
            <div class="filetop">
              <!-- <label for="filepicker">Choose a file</label> -->
              <input
                type="file"
                name="files"
                id="filepicker"
                multiple
                placeholder="Drop File"
              />
            </div>
            <button type="submit" id="upload" method="POST" class="fileupload">
              Upload
            </button>
          </form>
        </div>
        <div id="submitdiv">
          <form id="submitform" ,onsubmit="submitFile(event)">
            <label for="lva">Lehrveranstaltung:</label>
            <div class="autocomplete">
              <input
                type="text"
                id="lva"
                name="lva"
                placeholder="Lehrveranstaltung"
                autocomplete="off"
              />
            </div>
            <br />
            <!-- <br /> -->
            <label for="prof">Vortragende*r:</label>
            <div class="autocomplete">
              <input
                type="text"
                id="prof"
                name="prof"
                placeholder="Vortragende*r"
                autocomplete="off"
              />
            </div>
            <br />
            <!-- <br /> -->
            <label for="name">Name:</label>
            <input
              type="text"
              id="name"
              name="fname"
              placeholder="Prüfung"
            /><br />
            <label for="sem">Semester:</label>
            <input type="text" id="sem" name="sem" placeholder="2024W" /><br />
            <input
              type="radio"
              id="pruefung"
              name="stype"
              value="0"
              checked="checked"
            />
            <label for="pruefung">Prüfung</label><br />
            <input type="radio" id="klausur" name="stype" value="1" />
            <label for="klausur">Klausur</label><br />
            <input type="radio" id="uebung" name="stype" value="2" />
            <label for="uebung">Übung</label><br />
            <input type="radio" id="labor" name="stype" value="3" />
            <label for="labor">Labor</label><br />
            <input type="radio" id="unterlagen" name="stype" value="4" />
            <label for="unterlagen">Unterlagen</label><br />
            <input type="radio" id="zusammenfassungen" name="stype" value="5" />
            <label for="zusammenfassungen">Zusammenfassung</label><br />
            <input type="radio" id="multimedia" name="stype" value="6" />
            <label for="multimedia">Multimedia</label><br />
            <br />
            <div id="subcatdiv">
              <label for="subcat">Veranstaltung</label>
              <div class="autocomplete">
                <input
                  type="text"
                  id="subcat"
                  name="subcat"
                  placeholder="Klausur 1"
                  autocomplete="off"
                />
              </div>
            </div>
            <div id="datediv">
              <label for="date">Datum</label>
              <input
                type="date"
                id="date"
                name="ex_date"
                placeholder="Drop File"
              /><br />
            </div>
            <input
              type="checkbox"
              name="ocr"
              id="sec_censor"
              value="True"
            /><label for="sec_censor">OCR</label><br /><br />
            <button type="submit" id="send">Senden</button>
          </form>
        </div>
      </div>
      <div class="right" id="rightdiv">
        <div class="buttons" id="buttonsdiv">
          <button id="prev">Prev</button><button id="next">Next</button>
          <div>
            <span id="npage"></span>
            <span>/</span>
            <span id="npages"></span>
          </div>
          <button id="clr">Clear Page</button><button id="ca">Claer All</button>
        </div>
        <div id="cnvdiv">
          <div class="stack" id="cnvcont">
            <canvas id="cnv"></canvas>
            <canvas id="drw_cnv"></canvas>
          </div>
        </div>
      </div>
    </div>
  </body>
 </html>
--- a/app/init.py
+++ b/app/init.py
@@ -1,10 +1,17 @@
-import paramiko
+from os.path import isdir
 from stat import S_ISDIR, S_ISREG
 import re
 import pathlib
 import os
 # from base64 import decodebytes
 import json
 import mariadb
 import logging
 from pathlib import Path
 import schedule
 import time
 import pytz
 CATEGORIES = [
    "Prüfungen",
@@ -16,9 +23,24 @@ CATEGORIES = [
    "Multimedia",
 ]
 SUBCAT_CATEGORIES = ["Klausuren", "Übungen", "Labore"]
-unizeug_path = "/mnt/save/daten/Unizeug/"
+unizeug_path = os.environ.get("UNIZEUG_PATH", "./unizeug")
 APP_ROOT_PATH = Path(os.environ.get("APP_ROOT_PATH", "./app"))
 FILES_IN_PROGRESS = APP_ROOT_PATH / "files/"
 log = logging.getLogger(__name__)
 logging.basicConfig(
    filename="init.log",
    level=logging.INFO,
    format="[%(asctime)s, %(filename)s:%(lineno)s -> %(funcName)10s() ]%(levelname)s: %(message)s",
 )
 debug = log.debug
 info = log.info
 error = log.error
 db = mariadb.connect(
-    host="localhost", user="wildserver", password="DBPassword", database="Unizeug"
+    host=os.environ.get("DB_HOST", "db"),
    user=os.environ.get("DB_USER", "user"),
    password=os.environ.get("DB_PASSWORD", "DBPASSWORD"),
    database=os.environ.get("DB_DATABASE", "unizeug"),
 )
 c = db.cursor()
 try:
@@ -49,33 +71,54 @@ except mariadb.OperationalError:
 c.execute(
    "CREATE TABLE SubCats(id BIGINT(20) UNSIGNED NOT NULL AUTO_INCREMENT,LId BIGINT(20),PId BIGINT(20),cat TINYINT UNSIGNED,name VARCHAR(256), PRIMARY KEY(id))"
 )
 try:
    c.execute(
        "CREATE TABLE FIP(id UUID DEFAULT(UUID()), filename VARCHAR(256), filetype VARCHAR(8),initTimeStamp DATETIME, PRIMARY KEY(id))"
    )
 except mariadb.OperationalError:
    pass
 db.commit()
 def remove_old_FIP_entrys():
    cur = db.cursor(dictionary=True)
    cur.execute(
        "SELECT id,filename FROM FIP WHERE HOUR(TIMEDIFF(NOW(),initTimeStamp)) > 24 "
    )
    files = cur.fetchall()
    info(f"Remove Files: {files}")
    for file in files:
        c.execute("DELETE FROM FIP WHERE id=?", (file["id"],))
        os.remove(FILES_IN_PROGRESS / file["filename"])
    db.commit()
 def get_dirstruct():
-    with open("app/pwfile.json", "r") as f:
+    # with open("app/pwfile.json", "r") as f:
-        cred = json.load(f)
+    # cred = json.load(f)
-    ssh = paramiko.SSHClient()
+    # ssh = paramiko.SSHClient()
-    print(cred["sftpurl"])
+    # print(cred["sftpurl"])
-    ssh.set_missing_host_key_policy(paramiko.AutoAddPolicy())
+    # ssh.set_missing_host_key_policy(paramiko.AutoAddPolicy())
    # key=paramiko.RSAKey(data=decodebytes(bytes(cred["key"],"utf-8")))
    # ssh.get_host_keys().add(cred["sftpurl"], 'ssh-rsa', key)
-    ssh.connect(cred["sftpurl"], username=cred["sftpuser"], password=cred["sftpPW"])
+    # ssh.connect(cred["sftpurl"], username=cred["sftpuser"], password=cred["sftpPW"])
-    sftp = ssh.open_sftp()
+    # sftp = ssh.open_sftp()
-    folders = sftp.listdir_attr(unizeug_path)
+    # folders = sftp.listdir_attr(unizeug_path)
-    for entry in folders:
+    folders = pathlib.Path(unizeug_path)
    for entry in folders.iterdir():
        if entry is None:
            continue
-        if not S_ISDIR(entry.st_mode):
+        if not entry.is_dir():
            continue
-        fname = str(entry.filename)
+        fname = str(entry.name)
        regex = re.compile(r"Multimedia_only")
        if regex.search(fname):
            continue
        # print(fname)
        lvid = re.search(r"[a-zA-Z0-9]{3}\.[a-zA-Z0-9]{3}", fname)
-        print(lvid)
+        # print(lvid)
        if lvid is None:
            error(f"Didnt Find LVA ID in Directory {fname}")
            continue
        lvid = lvid.group()[:3] + lvid.group()[4:]
        # name = fname[:-8]
@@ -89,41 +132,38 @@ def get_dirstruct():
        cur.execute("SELECT id FROM LVAs WHERE lvid=?", (lvid,))
        lid = cur.fetchone()[0]
        db.commit()
-        for profsdir in sftp.listdir_attr(unizeug_path + fname + "/"):
+        for profsdir in entry.iterdir():
-            if profsdir is None or not S_ISDIR(profsdir.st_mode):
+            if profsdir is None:
                continue
            if not profsdir.is_dir():
                continue
            # print(profsdir.filename)
            try:
-                lastname, firstname = re.split(r"[_\-\s]", str(profsdir.filename))
+                lastname, firstname = re.split(r"[_\-\s]", str(profsdir.name))
                pid = link_prof(firstname, lastname, lid)
            except ValueError:
-                print(f"{name} is broken")
+                error(f"Couldnt get Profs from {fname}")
                continue
-            for cat in sftp.listdir_attr(
+            for cat in profsdir.iterdir():
-                unizeug_path + fname + "/" + profsdir.filename + "/"
+                if cat is None:
            ):
                if cat is None or not S_ISDIR(cat.st_mode):
                    continue
-                if cat.filename not in SUBCAT_CATEGORIES:
+                if not cat.is_dir():
                    continue
-                idx = CATEGORIES.index(cat.filename)
+                if cat.name not in SUBCAT_CATEGORIES:
-                for subcat in sftp.listdir_attr(
+                    continue
-                    unizeug_path
+                idx = CATEGORIES.index(cat.name)
-                    + fname
+                for subcat in cat.iterdir():
-                    + "/"
+                    if subcat is None:
-                    + profsdir.filename
+                        continue
-                    + "/"
+                    if not subcat.is_dir():
                    + cat.filename
                    + "/"
                ):
                    if subcat is None or not S_ISDIR(subcat.st_mode):
                        continue
                    cur = db.cursor()
                    cur.execute(
                        "INSERT INTO SubCats (LId,PId,cat,name) VALUES(?,?,?,?)",
-                        (lid, pid, idx, subcat.filename),
+                        (lid, pid, idx, subcat.name),
                    )
                db.commit()
    remove_old_FIP_entrys()
 def link_prof(firstname, lastname, lid):
@@ -150,3 +190,8 @@ def link_prof(firstname, lastname, lid):
 if __name__ == "__main__":
    get_dirstruct()
    info("Database updated")
    schedule.every().day.at("04:00", "Europe/Vienna").do(get_dirstruct)
    while True:
        schedule.run_pending()
        time.sleep(1)
--- a/app/init_ssh.py
+++ b/app/init_ssh.py
@@ -0,0 +1,152 @@
 import paramiko
 from stat import S_ISDIR, S_ISREG
 import re
 # from base64 import decodebytes
 import json
 import mariadb
 CATEGORIES = [
    "Prüfungen",
    "Klausuren",
    "Übungen",
    "Labore",
    "Unterlagen",
    "Zusammenfassungen",
    "Multimedia",
 ]
 SUBCAT_CATEGORIES = ["Klausuren", "Übungen", "Labore"]
 unizeug_path = "/mnt/save/daten/Unizeug/"
 db = mariadb.connect(
    host="localhost", user="wildserver", password="DBPassword", database="Unizeug"
 )
 c = db.cursor()
 try:
    c.execute("DROP TABLE LVAs")
 except mariadb.OperationalError:
    pass
 c.execute(
    "CREATE TABLE LVAs(id BIGINT(20) unsigned NOT NULL AUTO_INCREMENT,lvid VARCHAR(6), lvname VARCHAR(256), lvpath VARCHAR(256),PRIMARY KEY(id))"
 )
 try:
    c.execute("DROP TABLE Profs")
 except mariadb.OperationalError:
    pass
 c.execute(
    "CREATE TABLE Profs(id BIGINT(20) unsigned NOT NULL AUTO_INCREMENT,name VARCHAR(256),PRIMARY KEY(id))"
 )
 try:
    c.execute("DROP TABLE LPLink")
 except mariadb.OperationalError:
    pass
 c.execute(
    "CREATE TABLE LPLink(id BIGINT(20) unsigned NOT NULL AUTO_INCREMENT,LId bigint(20),PId bigint(20),PRIMARY KEY(id))"
 )
 try:
    c.execute("DROP TABLE SubCats")
 except mariadb.OperationalError:
    pass
 c.execute(
    "CREATE TABLE SubCats(id BIGINT(20) UNSIGNED NOT NULL AUTO_INCREMENT,LId BIGINT(20),PId BIGINT(20),cat TINYINT UNSIGNED,name VARCHAR(256), PRIMARY KEY(id))"
 )
 db.commit()
 def get_dirstruct():
    with open("app/pwfile.json", "r") as f:
        cred = json.load(f)
    ssh = paramiko.SSHClient()
    print(cred["sftpurl"])
    ssh.set_missing_host_key_policy(paramiko.AutoAddPolicy())
    # key=paramiko.RSAKey(data=decodebytes(bytes(cred["key"],"utf-8")))
    # ssh.get_host_keys().add(cred["sftpurl"], 'ssh-rsa', key)
    ssh.connect(cred["sftpurl"], username=cred["sftpuser"], password=cred["sftpPW"])
    sftp = ssh.open_sftp()
    folders = sftp.listdir_attr(unizeug_path)
    for entry in folders:
        if entry is None:
            continue
        if not S_ISDIR(entry.st_mode):
            continue
        fname = str(entry.filename)
        regex = re.compile(r"Multimedia_only")
        if regex.search(fname):
            continue
        # print(fname)
        lvid = re.search(r"[a-zA-Z0-9]{3}\.[a-zA-Z0-9]{3}", fname)
        print(lvid)
        if lvid is None:
            continue
        lvid = lvid.group()[:3] + lvid.group()[4:]
        # name = fname[:-8]
        name = re.sub(r"[a-zA-Z0-9]{3}\.[a-zA-Z0-9]{3}", "", fname)
        # print(name)
        # print(lvid)
        cur = db.cursor()
        cur.execute(
            "INSERT INTO LVAs (lvid, lvname, lvpath) VALUES(?,?,?)", (lvid, name, fname)
        )
        cur.execute("SELECT id FROM LVAs WHERE lvid=?", (lvid,))
        lid = cur.fetchone()[0]
        db.commit()
        for profsdir in sftp.listdir_attr(unizeug_path + fname + "/"):
            if profsdir is None or not S_ISDIR(profsdir.st_mode):
                continue
            # print(profsdir.filename)
            try:
                lastname, firstname = re.split(r"[_\-\s]", str(profsdir.filename))
                pid = link_prof(firstname, lastname, lid)
            except ValueError:
                print(f"{name} is broken")
                continue
            for cat in sftp.listdir_attr(
                unizeug_path + fname + "/" + profsdir.filename + "/"
            ):
                if cat is None or not S_ISDIR(cat.st_mode):
                    continue
                if cat.filename not in SUBCAT_CATEGORIES:
                    continue
                idx = CATEGORIES.index(cat.filename)
                for subcat in sftp.listdir_attr(
                    unizeug_path
                    + fname
                    + "/"
                    + profsdir.filename
                    + "/"
                    + cat.filename
                    + "/"
                ):
                    if subcat is None or not S_ISDIR(subcat.st_mode):
                        continue
                    cur = db.cursor()
                    cur.execute(
                        "INSERT INTO SubCats (LId,PId,cat,name) VALUES(?,?,?,?)",
                        (lid, pid, idx, subcat.filename),
                    )
                db.commit()
 def link_prof(firstname, lastname, lid):
    cur = db.cursor()
    cur.execute("SELECT id from Profs WHERE name=?", (lastname + " " + firstname,))
    res = cur.fetchone()
    if res is not None:
        cur.execute("INSERT INTO LPLink (LId,PId) VALUES(?,?)", (lid, res[0]))
        db.commit()
        return res[0]
    cur.execute("SELECT id from Profs WHERE name=?", (firstname + " " + lastname,))
    res = cur.fetchone()
    if res is not None:
        cur.execute("INSERT INTO LPLink (LId,PId) VALUES(?,?)", (lid, res[0]))
        db.commit()
        return res[0]
    cur.execute("INSERT INTO Profs (name) VALUES(?)", (lastname + " " + firstname,))
    cur.execute("SELECT id FROM Profs WHERE name=?", (lastname + " " + firstname,))
    res = cur.fetchone()
    cur.execute("INSERT INTO LPLink (LId,PId) VALUES(?,?)", (lid, res[0]))
    db.commit()
    return res[0]
 if __name__ == "__main__":
    get_dirstruct()
--- a/app/main.py
+++ b/app/main.py
@@ -1,44 +1,62 @@
 from typing import Annotated
 from typing import List, Dict, Tuple, Sequence
 from starlette.responses import StreamingResponse
 from annotated_types import IsDigit
 from fastapi import FastAPI, File, HTTPException, UploadFile, Request, Form
 from fastapi.responses import FileResponse
 # import multiprocessing
 # import threading
 # import concurrent.futures
-# import asyncio
+import asyncio
 # import fastapi
 from fastapi.staticfiles import StaticFiles
 import pymupdf
 # import fitz as pymupdf
 import json
 import re
 import os
 import signal
 import mariadb
 import sys
 import filetype
 import datetime
 import logging
-import inspect
+import pathlib
 from pathlib import Path
 from starlette.types import HTTPExceptionHandler
 log = logging.getLogger(__name__)
 logging.basicConfig(
-    filename="app.log",
+    filename=os.environ.get("APP_LOG_PATH"),
    level=logging.INFO,
-    format="[%(asctime)s, %(filename)s:%(lineno)s -> %(funcName)10s() ]%(levelname)s: %(message)s",
+    format="[%(asctime)s, %(filename)s:%(lineno)s -> %(funcName)10s()] %(levelname)s: %(message)s",
 )
 debug = log.debug
 info = log.info
 error = log.error
 critical = log.critical
 def exception_handler(etype, value, tb):
    log.exception(f"Uncought Exception: {value}")
 sys.excepthook = exception_handler
 db = mariadb.connect(
-    host="localhost", user="wildserver", password="DBPassword", database="Unizeug"
+    host=os.environ.get("DB_HOST", "db"),
    user=os.environ.get("DB_USER", "user"),
    password=os.environ.get("DB_PASSWORD", "DBPASSWORD"),
    database=os.environ.get("DB_DATABASE", "unizeug"),
 )
 info("App Started")
@@ -47,8 +65,6 @@ info("App Started")
 # startup()
 app = FastAPI()
 app.mount("/favicon", StaticFiles(directory="./favicon"), name="favicon")
 app.mount("/static", StaticFiles(directory="./static"), name="static")
 CATEGORIES = [
@@ -60,15 +76,18 @@ CATEGORIES = [
    "Zusammenfassungen",
    "Multimedia",
 ]
 APP_ROOT_PATH = Path(os.environ.get("APP_ROOT_PATH", "./app"))
 SUBCAT_CATEGORIES = ["Klausuren", "Übungen", "Labore"]
 SUBCAT_CATEGORIES_I = [1, 2, 3]
 EX_DATE_CATEGORIES = ["Prüfungen", "Klausuren"]
 EX_DATE_CATEGORIES_I = [0, 1]
-UNIZEUG_PATH = "./app/dest/"
+UNIZEUG_PATH = Path(os.environ.get("UNIZEUG_PATH", "./app/dest"))
-FILES_IN_PROGRESS = "./app/files/"
+FILES_IN_PROGRESS = APP_ROOT_PATH / "files/"
-EMPTYFILE = "./app/graphics/empty.pdf"
+EMPTYFILE = APP_ROOT_PATH / "graphics/empty.pdf"
-UNSUPPORTEDFILE = "./app/graphics/unsupported.pdf"
+UNSUPPORTEDFILE = APP_ROOT_PATH / "graphics/unsupported.pdf"
-GREETINGFILE = "./app/graphics/greeting.pdf"
+GREETINGFILE = APP_ROOT_PATH / "graphics/greeting.pdf"
 FAVICON = APP_ROOT_PATH / "favicon"
 STATIC_FILES = APP_ROOT_PATH / "static"
 # cur = db.cursor()
@@ -76,6 +95,9 @@ GREETINGFILE = "./app/graphics/greeting.pdf"
 # for l in cur:
 #     print(l)
 # locpaths = ["./VO_Mathematik_3.pdf"]  # replace this with a database
 censor_status_update_events: Dict[str, asyncio.Event] = {}
 censor_status_datas: Dict[str, Dict[str, int | None | str | bool]] = {}
 # censor_finished_flags: Dict[str, asyncio.Event] = {}
 def _sql_quarry(
@@ -105,12 +127,40 @@ def _sql_quarry(
        )
 def sql_connector_is_active(connector: mariadb.Connection) -> bool:
    try:
        connector.ping()
    except mariadb.Error as e:
        return False
    return True
 def sql_connect(connector: mariadb.Connection) -> mariadb.Connection:
    try:
        connector = mariadb.connect(
            host=os.environ.get("DB_HOST", "db"),
            user=os.environ.get("DB_USER", "user"),
            password=os.environ.get("DB_PASSWORD", "DBPASSWORD"),
            database=os.environ.get("DB_DATABASE", "Unizeug"),
        )
    except mariadb.Error as e:
        critical(
            f"Cannot reconnect to Database {os.environ.get('DB_DATABASE', 'Unizeug')} on {os.environ.get('DB_HOST', 'db')}. Got Mariadb Error: {e}"
        )
        os.kill(os.getpid(), signal.SIGTERM)
        raise HTTPException(500, detail="Database failed")
    return connector
 def sql(
    querry: str,
    data: Tuple[str | int, ...] | str | int = (),
    return_result: bool = True,
    commit: bool = False,
 ) -> List[Tuple]:
    global db
    if not sql_connector_is_active(db):
        db = sql_connect(db)
    cur = db.cursor(dictionary=False)
    return _sql_quarry(cur, querry, data, return_result, commit)
@@ -121,6 +171,10 @@ def sqlT(
    return_result: bool = True,
    commit: bool = False,
 ) -> List[Dict]:
    global db
    if not sql_connector_is_active(db):
        db = sql_connect(db)
    cur = db.cursor(dictionary=True)
    return _sql_quarry(cur, querry, data, return_result, commit)
@@ -139,10 +193,22 @@ def sqlT(
    #     )
 app.mount(
    "/favicon",
    StaticFiles(directory=os.environ.get("FAVICON_PATH", FAVICON)),
    name="favicon",
 )
 app.mount(
    "/static",
    StaticFiles(directory=os.environ.get("STATIC_PATH", STATIC_FILES)),
    name="static",
 )
@app.get("/")
 async def get_index():
    """gives the Index.html file"""
-    return FileResponse("./index.html")
+    return FileResponse(APP_ROOT_PATH / "index.html")
@app.get("/files/{file_id}")
@@ -169,7 +235,7 @@ async def get_file(file_id: str):
    # status_code=500, detail="Somethings wrong with the database"
    # )
    # filename = cur.fetchone()[0]
-    return FileResponse(FILES_IN_PROGRESS + filename)
+    return FileResponse(FILES_IN_PROGRESS / filename)
@app.get("/search/lva")
@@ -215,6 +281,9 @@ async def search_lva(
            )
            # res += cur.fetchall()
        res = remove_duplicates(res + zw)
    info(
        f"LVA Search: {searchterm}; Result: {res[: (searchlim if searchlim != 0 else -1)]}"
    )
    if searchlim == 0:
        return res
    else:
@@ -249,6 +318,9 @@ async def search_profs(
        )
        # res += cur.fetchall()
    res = remove_duplicates(res + zw)
    info(
        f"Prof Search: {searchterm}; Result: {res[: (searchlim if searchlim != 0 else -1)]}"
    )
    if searchlim == 0:
        return res
    else:
@@ -289,6 +361,9 @@ async def search_subcats(
        )
        # res += cur.fetchall()
    res = remove_duplicates(res + rest)
    info(
        f"Subcatrgory Search: {searchterm}; Result: {res[: (searchlim if searchlim != 0 else -1)]}"
    )
    if searchlim == 0:
        return res
    else:
@@ -345,7 +420,7 @@ async def create_upload_file(files: List[UploadFile], c2pdf: bool = True):
            content = doc.tobytes()
    if ft != "dir":
        filename = make_filename_unique(filename)
-    locpath = FILES_IN_PROGRESS + filename
+    locpath = FILES_IN_PROGRESS / filename
    # locpaths.append(locpath)
    # cur = db.cursor()
    # try:
@@ -405,14 +480,14 @@ async def get_submission(
    pagescales: Annotated[
        str, Form()
    ],  # Scales of Pages  # Annotated[List[Dict[str, float]], Form()],
-    censor: Annotated[str, Form()],
+    ocr: Annotated[str, Form()],
 ):
    """handles submission"""
    print(
-        f"lva: {lva}, prof: {prof}, fname {fname}, stype: {stype}, subcat: {subcat}, sem: {sem}, ex_date: {ex_date}, rects: {rects}, pagescales: {pagescales}, censor: {censor}"
+        f"lva: {lva}, prof: {prof}, fname {fname}, stype: {stype}, subcat: {subcat}, sem: {sem}, ex_date: {ex_date}, rects: {rects}, pagescales: {pagescales}, ocr: {ocr}"
    )
    info(
-        f"lva: {lva}, prof: {prof}, fname {fname}, stype: {stype}, subcat: {subcat}, sem: {sem}, ex_date: {ex_date}, rects: {rects}, pagescales: {pagescales}, censor: {censor}"
+        f"Got Submission: lva: {lva}, prof: {prof}, fname {fname}, stype: {stype}, subcat: {subcat}, sem: {sem}, ex_date: {ex_date}, rects: {rects}, pagescales: {pagescales}, ocr: {ocr}"
    )
    rects_p = json.loads(rects)
    scales_p = json.loads(pagescales)
@@ -429,7 +504,7 @@ async def get_submission(
            error(f"User tried to upload a file without specifying the {th[1]}")
            raise HTTPException(400, f"You need to specify a {th[1]}")
-    filepath = "./app/files/" + res[0][0]
+    filepath = FILES_IN_PROGRESS / res[0][0]
    # except mariadb.Error as e:
    # print(f"Mariadb Error: {e}")
    # raise HTTPException(
@@ -441,25 +516,73 @@ async def get_submission(
    except ValueError as e:
        error(f"Error creating savepath: {e}")
        raise HTTPException(status_code=400, detail=f"Error creation savepath: {e}")
-    await censor_pdf(
+    # censor_finished_flags[fileId] = asyncio.Event()
-        filepath, dest, rects_p, scales_p, False if censor == "False" else True
+    censor_status_datas[fileId] = {}
-    )
+    if fileId not in censor_status_update_events:
        censor_status_update_events[fileId] = asyncio.Event()
    if ocr == "True":
        await asyncio.to_thread(
            censor_pdf_ocr,
            filepath,
            dest,
            rects_p,
            scales_p,
            fileId,
        )
    else:
        await asyncio.to_thread(
            censor_pdf,
            filepath,
            dest,
            rects_p,
            scales_p,
            fileId,
        )
    # return {"done": "ok"}
    # print(dest)
    # await censor_finished_flags[fileId].wait()
    # censor_finished_flags[fileId].clear()
    info(f"Saved file {fileId} as {dest}")
    delete_from_FIP(fileId)
    return FileResponse(dest, content_disposition_type="inline")
-async def censor_pdf(
+@app.get("/get_censor_status/{file_id}")
-    path: str,
+async def get_censor_status(file_id: str):
-    destpath: str,
+    """Yields the currrent page being censored and the total number of pages"""
    if len(sql("Select filename from FIP where id=?", (file_id,))) < 1:
        raise HTTPException(
            400,
            detail="You are trying to get a status updater for a file that dosent exist.",
        )
    if file_id not in censor_status_update_events:
        censor_status_update_events[file_id] = asyncio.Event()
    return StreamingResponse(
        yield_censor_status(file_id), media_type="text/event-stream"
    )
 async def yield_censor_status(file_id: str):
    """Internal function to yield updates to the stream"""
    while True:
        await censor_status_update_events[file_id].wait()
        censor_status_update_events[file_id].clear()
        yield f"event: censorUpdate\ndata: {json.dumps(censor_status_datas[file_id])}\n\n"
        if censor_status_datas[file_id]["done"]:
            del censor_status_update_events[file_id]
            del censor_status_datas[file_id]
            return
 def censor_pdf(
    path: os.PathLike,
    destpath: os.PathLike,
    rects: List[List[List[float]]],
    scales: List[Dict[str, float]],
-    secure: bool,
+    file_id: str,
 ):
-    """Censors pdf and runs OCR
+    """Censors pdf and saves the file to the given Destpath.
    If Secure is True the file is converted to Pixels and then recreated; else the censored sections are just covering the text below and can be easiliy removed with e.g. Inkscape
    Args:
        path: path to the pdf document
        destpath: Path where the result is supposed to be saved to
@@ -469,15 +592,62 @@ async def censor_pdf(
    Returns:
        None
    """
    info(f"started Censoring for file {path} to be saved to {destpath}")
    doc = pymupdf.open(path)
    page = doc[0]
    npage = doc.page_count
    for i in range(npage):
        page = doc[i]
        if i < len(rects) and rects[i] != []:
            print(i)
            wfac = page.rect.width / scales[i]["width"]
            hfac = page.rect.height / scales[i]["height"]
            for rect in rects[i]:
                prect = pymupdf.Rect(
                    rect[0] * wfac,
                    rect[1] * hfac,
                    (rect[0] + rect[2]) * wfac,
                    (rect[1] + rect[3]) * hfac,
                )
                page.add_redact_annot(
                    prect,
                    fill=(0, 0, 0),
                )
            page.apply_redactions()
        censor_status_datas[file_id]["page"] = i + 1
        censor_status_datas[file_id]["pages"] = npage
        censor_status_datas[file_id]["done"] = False
        censor_status_update_events[file_id].set()
    doc.set_metadata({})
    doc.save(destpath, garbage=4, deflate=True, clean=True)
    censor_status_datas[file_id]["done"] = True
    censor_status_update_events[file_id].set()
 def censor_pdf_ocr(
    path: os.PathLike,
    destpath: os.PathLike,
    rects: List[List[List[float]]],
    scales: List[Dict[str, float]],
    file_id: str,
 ):
    """Censors pdf and runs OCR
    The file is converted to Pixels and then recreated.
    Saves the file to the given Destpath.
    Args:
        path: path to the pdf document
        destpath: Path where the result is supposed to be saved to
        rects: Coordinates of rectangles to be placed on the pdf document
        scales: Scales of the rects coordinates for the pdf document
        secure: weather or not the pdf document is supposed to be converted into an Image (and back) to make shure, the censoring is irreversible
    Returns:
        None
    """
    info(f"started Censoring in OCR Mode for file {path} to be saved to {destpath}")
    doc = pymupdf.open(path)
    output = pymupdf.open()
    page = doc[0]
    # width = page.rect.width
    # height = page.rect.height
    # print(width, height)
    npage = doc.page_count
    # pages = []
    # tasks = []
    for i in range(npage):
        page = doc[i]
        if i < len(rects) and rects[i] != []:
@@ -496,38 +666,41 @@ async def censor_pdf(
                    color=(0, 0, 0),
                    fill=(0, 0, 0),
                )
-        if secure:
+        censor_status_datas[file_id]["page"] = i + 1
-            # pages.append(page)
+        censor_status_datas[file_id]["pages"] = npage
        censor_status_datas[file_id]["done"] = False
        censor_status_update_events[file_id].set()
        # THis Costs us dearly
        try:
            bitmap = page.get_pixmap(dpi=400)
            pdf_bytes = bitmap.pdfocr_tobytes(
                language="deu",
                tessdata="/usr/share/tessdata/",  # tesseract needs to be installed; this is the path to thetesseract files
            )
            output.insert_pdf(pymupdf.Document(stream=pdf_bytes))
-            # tasks.append(asyncio.create_task(censor_page(page)))
+        except RuntimeError as e:
-            print(f"Page {i + 1}/{npage}: CENSORING DONE")
+            error(
-        else:
+                f"Error in OCR for document: {destpath}. Error: {e}. Falling back to standard mode."
-            output.insert_pdf(doc, i, i)
+            )
-
+            if i < len(rects) and rects[i] != []:
-    # if secure:
+                for rect in rects[i]:
-    # pages_bytes: List[bytes] = []
+                    prect = pymupdf.Rect(
-    # censor_page(pages[0])
+                        rect[0] * wfac,
-    # with multiprocessing.Pool(npage) as p:
+                        rect[1] * hfac,
-    # pages_bytes = p.map(censor_page, pages)
+                        (rect[0] + rect[2]) * wfac,
-    # pages_bytes = p.map(test_function, [1, 2, 3, 4])
+                        (rect[1] + rect[3]) * hfac,
-    # for pdf_bytes in pages_bytes:
+                    )
-    # output.insert_pdf(pymupdf.Document(stream=pdf_bytes))
+                    page.add_redact_annot(
-    # with concurrent.futures.ThreadPoolExecutor() as executor:
+                        prect,
-    #     futures = []
+                        fill=(0, 0, 0),
-    #     for page in pages:
+                    )
-    #         futures.append(executor.submit(censor_page, page))
+                page.apply_redactions()
-    #     for future in futures:
+            output.insert_pdf(page.parent, from_page=page.number, to_page=page.number)
-    #         output.insert_pdf(pymupdf.Document(stream=future.result()))
+        # End of the costly part
-    #
+        print(f"Page {i + 1}/{npage}: CENSORING DONE")
    # for task in tasks:
    # output.insert_pdf(pymupdf.Document(stream=await task))
    # print("CENSORING DONE")
    output.save(destpath)
    censor_status_datas[file_id]["done"] = True
    censor_status_update_events[file_id].set()
 def test_function(i: int) -> bytes:
@@ -579,21 +752,22 @@ def make_savepath(
    ex_date: str,
    fname: str,
    ftype: str,
-) -> str:
+) -> os.PathLike:
    """Generates the path, the file is saved to after the upload process is finished. It creates all nessecery directories."""
    info(f"Started to make Savepath for '{fname}' in '{lva}' with prof '{prof}'.")
    lv = get_lvpath(lva)
-    lvpath = lv[1] + "/"
+    lvpath = Path(lv[1])
    pf = get_profpath(prof, lv[0])
-    pfpath = pf[1] + "/"
+    pfpath = Path(pf[1])
-    catpath = CATEGORIES[int(cat)] + "/"
+    catpath = Path(CATEGORIES[int(cat)])
-    scpath = ""
+    scpath: str | os.PathLike = ""
    if int(cat) in SUBCAT_CATEGORIES_I and subcat != "":
        sc = get_subcatpath(subcat, int(cat), pf[0], lv[0])
-        scpath = sc[1] + "/"
+        scpath = Path(sc[1])
    if int(cat) == 6:
-        savepath = UNIZEUG_PATH + lv[1] + "_Multimedia_only/" + pfpath
+        savepath = UNIZEUG_PATH / (lv[1] + "_Multimedia_only/") / pfpath
    else:
-        savepath = UNIZEUG_PATH + lvpath + pfpath + catpath + scpath
+        savepath = UNIZEUG_PATH / lvpath / pfpath / catpath / scpath
    os.makedirs(savepath, exist_ok=True)
    filename = sem + "_"
    if int(cat) in EX_DATE_CATEGORIES_I:
@@ -607,9 +781,20 @@ def make_savepath(
                400,
                "You have not specified a date for an upload that requires a date like an exam.",
            )
-        filename += yyyy + "_" + mm + "_" + dd + "_"
+        filename = yyyy + "_" + mm + "_" + dd + "_"
-    filename += fname + "." + ftype
+    filename += fname
-    return savepath + filename
+    file = filename + "." + ftype
    destpath = savepath / file
    i = 0
    while destpath.is_file():
        info(f"{destpath} already exists.")
        file = filename + f"_{i}." + ftype
        i += 1
        destpath = savepath / file
    destpath.touch()
    info(f"Path for file to be saved generated as: {savepath / file}")
    return savepath / file
 def get_lvpath(lva: str) -> Tuple[int, str]:
@@ -802,10 +987,10 @@ async def save_files_to_folder(files: List[UploadFile]) -> str:
    if filename == "":
        filename = "None"
    filename = make_filename_unique(filename)
-    os.mkdir(FILES_IN_PROGRESS + filename)
+    os.mkdir(FILES_IN_PROGRESS / filename)
    for idx, file in enumerate(files):
        fn = file.filename if file.filename is not None else "None" + str(idx)
-        with open(FILES_IN_PROGRESS + filename + "/" + fn, "wb") as f:
+        with open(FILES_IN_PROGRESS / filename / fn, "wb") as f:
            f.write(await file.read())
    return filename
@@ -833,13 +1018,13 @@ async def remove_old_FIP_entrys():
    info(f"Remove Files: {files}")
    for file in files:
        sql("DELETE FROM FIP WHERE id=?", (file["id"]), return_result=False)
-        os.remove(FILES_IN_PROGRESS + file["filename"])
+        os.remove(FILES_IN_PROGRESS / file["filename"])
    # sql(
    #     "DELETE FROM FIP WHERE HOUR(TIMEDIFF(NOW(),initTimeStamp)) > 24",
    #     return_result=False,
    # )
    db.commit()
-    return FileResponse("./index.html")
+    return FileResponse(APP_ROOT_PATH / "index.html")
 def delete_from_FIP(uuid: str):
@@ -847,4 +1032,4 @@ def delete_from_FIP(uuid: str):
    if len(res) < 1:
        raise HTTPException(500, "I am trying to delete a file that dose not exist")
    sql("DELETE FROM FIP WHERE id=?", (uuid,), return_result=False, commit=True)
-    os.remove(FILES_IN_PROGRESS + res[0]["filename"])
+    os.remove(FILES_IN_PROGRESS / res[0]["filename"])
--- a/app/static/app.js
+++ b/app/static/app.js
@@ -188,6 +188,9 @@ class PDFDocument {
  }
 }
 var mouseIsDown = false;
 var modal;
 var close_loading;
 var upload_status;
 //var startX = 0;
 //var startY = 0;
 //var pdf;
@@ -273,18 +276,39 @@ function submitPdf(eve) {
  formdata.append("fileId", doc.fID);
  //formdata.append("filename", doc.filename);
  formdata.append("ftype", doc.filetype);
-  if (!formdata.has("censor")) {
+  if (!formdata.has("ocr")) {
-    formdata.append("censor", "False");
+    formdata.append("ocr", "False");
  }
  console.log(formdata);
  submitForm(formdata);
 }
 async function submitForm(formData) {
  var updateEventSource = null;
  try {
-    const response = await fetch("http://127.0.0.1:8000/submit", {
+    updateEventSource = new EventSource("/get_censor_status/" + doc.fID);
    modal.style.display = "flex";
    // console.log("http://127.0.0.1:8000/get_censor_status/" + doc.fID);
    updateEventSource.addEventListener("censorUpdate", function(eve) {
      console.log(eve.data);
      var data = JSON.parse(eve.data);
      upload_status.innerText =
        "Censoring Page " + data.page + "/" + data.pages;
    });
  } catch {
    console.error(
      "Error geting eventsource for updating censoring page count: " + error,
    );
  }
  try {
    const response = await fetch("/submit/", {
      method: "POST",
      body: formData,
    });
    if (updateEventSource !== null) {
      updateEventSource.close();
    }
    modal.style.display = "none";
    //let responseJSON=await response.json();
    if (response.ok) {
      console.log("Submit OK");
@@ -304,7 +328,7 @@ async function submitForm(formData) {
      window.alert("Error: " + (await response.json())["detail"]);
    }
  } catch (error) {
-    console.error("Error" + error);
+    console.error("Error submitting: " + error);
  }
 }
 function uploadPdf(eve) {
@@ -322,7 +346,7 @@ function uploadPdf(eve) {
 }
 async function uploadFile(formData) {
  try {
-    const response = await fetch("http://127.0.0.1:8000/uploadfile", {
+    const response = await fetch("/uploadfile/", {
      method: "POST",
      body: formData,
    });
@@ -366,6 +390,14 @@ function initListeners() {
    doc.clearAll();
  });
 }
 function initLoading() {
  modal = document.querySelector("#loading");
  // close_loading = document.querySelector(".close");
  upload_status = document.querySelector("#upload_status");
  // close_loading.addEventListener("click", function() {
  // modal.style.display = "none";
  // });
 }
 const startPdf = () => {
  // doc = new PDFDocument(
  //   "./files/b78c869f-e0bb-11ef-9b58-84144d05d665",
@@ -374,6 +406,7 @@ const startPdf = () => {
  // );
  //pdf = new PDFView("./VO_Mathematik_3.pdf");
  doc = new PDFDocument("./files/greeting", "greeting", "pdf");
  initLoading();
  initDraw();
  initUpload();
  initListeners();
--- a/app/static/autocomplete.js
+++ b/app/static/autocomplete.js
@@ -1,4 +1,4 @@
-var url = "http://127.0.0.1:8000/search/";
+var url = "/search/";
 var lid = null;
 var pid = null;
 var activeAutocompletion = null;
@@ -21,7 +21,7 @@ function autocomplete(inp, type) {
      i,
      apirq,
      iname,
-      val = this.value;
+      val = inp.value;
    /*close any already open lists of autocompleted values*/
    closeAllLists();
    if (!val && type === "lva" && pid === null) {
@@ -56,7 +56,7 @@ function autocomplete(inp, type) {
    a.setAttribute("id", this.id + "autocomplete-list");
    a.setAttribute("class", "autocomplete-items");
    /*append the DIV element as a child of the autocomplete container:*/
-    this.parentNode.appendChild(a);
+    inp.parentNode.appendChild(a);
    /*for each item in the array...*/
    //await response;
    if (response.ok) {
--- a/app/static/dynhide.js
+++ b/app/static/dynhide.js
--- a/app/static/filedrop.js
+++ b/app/static/filedrop.js
--- a/app/static/style.css
+++ b/app/static/style.css
@@ -238,3 +238,84 @@ input[type="file"]::file-selector-button {
  width: 100%;
  /* background-color: purple; */
 }
 /* The Modal (background) */
 .modal {
  display: none;
  /* Hidden by default */
  position: fixed;
  /* Stay in place */
  z-index: 1;
  /* Sit on top */
  left: 0;
  top: 0;
  width: 100%;
  /* Full width */
  height: 100%;
  /* Full height */
  overflow: auto;
  /* Enable scroll if needed */
  background-color: #4f5977;
  /* Fallback color */
  background-color: rgba(0, 0, 0, 0.4);
  /* Black w/ opacity */
  justify-content: center;
 }
 /* Modal Content/Box */
 .loading-content {
  background-color: #4f5977;
  margin: auto;
  /* 15% from the top and centered */
  padding: 20px;
  /* border: 1px solid #888; */
  /* width: 80%; */
  border-radius: 15px;
  display: flex;
  flex-direction: column;
  /* Could be more or less, depending on screen size */
  align-items: center;
  text-align: center;
 }
 /* The Close Button */
 .close {
  color: #aaa;
  float: right;
  font-size: 28px;
  font-weight: bold;
 }
 .close:hover,
 .close:focus {
  color: black;
  text-decoration: none;
  cursor: pointer;
 }
 .upload_status_text {
  color: #ffffff;
  font-size: 16pt;
 }
 .loader {
  margin: auto;
  border: 16px solid #f3f3f3;
  /* Light grey */
  border-top: 16px solid #3498db;
  /* Blue */
  border-radius: 50%;
  width: 120px;
  height: 120px;
  animation: spin 2s linear infinite;
 }
@keyframes spin {
  0% {
    transform: rotate(0deg);
  }
  100% {
    transform: rotate(360deg);
  }
 }
--- a/compose.yml
+++ b/compose.yml
@@ -0,0 +1,67 @@
 services:
  app:
    container_name: python-app
    # command: python -m uvicorn app.main:app --host 0.0.0.0 --port 80
    build:
      context: .
      dockerfile: Dockerfile
    volumes:
      - ./app:/python
      - ./unizeug:/unizeug:source
    ports:
      - 80:80
    restart: unless-stopped
    environment:
      ENTRY_COMMAND: python -m uvicorn main:app --host 0.0.0.0 --port 80
      APP_LOG_PATH: /python/app.log
      APP_ROOT_PATH: /python
      UNIZEUG_PATH: /unizeug
      DB_HOST: db
      DB_USER: app
      DB_PASSWORD: DBPassword
      DB_DATABASE: Unizeug
      TZ: "Europe/Vienna"
    depends_on:
      - db
      - scaner
  db:
    container_name: db
    image: mariadb
    restart: unless-stopped
    environment:
      MARIADB_ROOT_PASSWORD: DBPassword
      MARIADB_USER: app
      UNIZEUG_PATH: /unizeug
      MARIADB_PASSWORD: DBPassword
      MARIADB_DATABASE: Unizeug
      TZ: "Europe/Vienna"
    healthcheck:
      test: ["CMD", "healthcheck.sh", "--connect", "--innodb_initialized"]
      start_period: 10s
      interval: 10s
      timeout: 5s
      retries: 3
    volumes:
      - ./mariadb:/var/lib/mysql
  scaner:
    container_name: python-scaner
    # command: python /python/init.py
    build:
      context: .
      dockerfile: Dockerfile
    volumes:
      - ./app:/python
      - ./unizeug:/unizeug:source
    restart: unless-stopped
    environment:
      ENTRY_COMMAND: python /python/init.py
      UNIZEUG_PATH: /unizeug
      APP_ROOT_PATH: /python
      DB_HOST: db
      DB_USER: app
      DB_PASSWORD: DBPassword
      DB_DATABASE: Unizeug
      TZ: "Europe/Vienna"
    depends_on:
      - db
--- a/entrypoint.sh
+++ b/entrypoint.sh
@@ -0,0 +1,4 @@
 #!/bin/sh 
 if [[ -n "$ENTRY_COMMAND" ]]; then
  /bin/sh -c "$ENTRY_COMMAND"
 fi
--- a/index.html
+++ b/index.html
@@ -1,103 +0,0 @@
 <!doctype html>
 <html lang="de">
 <head>
  <title>Unizeug uploader</title>
  <link rel="stylesheet" href="static/style.css" />
  <script src="https://cdnjs.cloudflare.com/ajax/libs/pdf.js/2.4.456/pdf.min.js"></script>
  <script src="static/app.js" defer></script>
  <script src="static/autocomplete.js" defer></script>
  <script src="static/dynhide.js" defer></script>
  <script src="static/filedrop.js" defer></script>
  <link rel="icon" type="image/png" href="/favicon/favicon-96x96.png" sizes="96x96" />
  <link rel="icon" type="image/svg+xml" href="/favicon/favicon.svg" />
  <link rel="shortcut icon" href="/favicon/favicon.ico" />
  <link rel="apple-touch-icon" sizes="180x180" href="/favicon/apple-touch-icon.png" />
  <meta name="apple-mobile-web-app-title" content="Unizeug" />
  <link rel="manifest" href="/favicon/site.webmanifest" />
 </head>
 <body>
  <div class="main">
    <div class="left" id="controldiv">
      <div id="fileupload">
        <form id="uploadform" enctype="multipart/form-data">
          <div class="filetop">
            <!-- <label for="filepicker">Choose a file</label> -->
            <input type="file" name="files" id="filepicker" multiple placeholder="Drop File" />
          </div>
          <button type="submit" id="upload" method="POST" class="fileupload">
            Upload
          </button>
        </form>
      </div>
      <div id="submitdiv">
        <form id="submitform" ,onsubmit="submitFile(event)">
          <label for="lva">Lehrveranstaltung:</label>
          <div class="autocomplete">
            <input type="text" id="lva" name="lva" placeholder="Lehrveranstaltung" autocomplete="off" />
          </div>
          <br />
          <!-- <br /> -->
          <label for="prof">Vortragende*r:</label>
          <div class="autocomplete">
            <input type="text" id="prof" name="prof" placeholder="Vortragende*r" autocomplete="off" />
          </div>
          <br />
          <!-- <br /> -->
          <label for="name">Name:</label>
          <input type="text" id="name" name="fname" placeholder="Prüfung" /><br />
          <label for="sem">Semester:</label>
          <input type="text" id="sem" name="sem" placeholder="2024W" /><br />
          <input type="radio" id="pruefung" name="stype" value="0" checked="checked" />
          <label for="pruefung">Prüfung</label><br />
          <input type="radio" id="klausur" name="stype" value="1" />
          <label for="klausur">Klausur</label><br />
          <input type="radio" id="uebung" name="stype" value="2" />
          <label for="uebung">Übung</label><br />
          <input type="radio" id="labor" name="stype" value="3" />
          <label for="labor">Labor</label><br />
          <input type="radio" id="unterlagen" name="stype" value="4" />
          <label for="unterlagen">Unterlagen</label><br />
          <input type="radio" id="zusammenfassungen" name="stype" value="5" />
          <label for="zusammenfassungen">Zusammenfassung</label><br />
          <input type="radio" id="multimedia" name="stype" value="6" />
          <label for="multimedia">Multimedia</label><br />
          <br />
          <div id="subcatdiv">
            <label for="subcat">Veranstaltung</label>
            <div class="autocomplete">
              <input type="text" id="subcat" name="subcat" placeholder="Klausur 1" autocomplete="off" />
            </div>
          </div>
          <div id="datediv">
            <label for="date">Datum</label>
            <input type="date" id="date" name="ex_date" placeholder="Drop File" /><br />
          </div>
          <input type="checkbox" name="censor" id="sec_censor" value="True" checked /><label
            for="sec_censor">Zensieren</label><br /><br />
          <button type="submit" id="send">Senden</button>
        </form>
      </div>
    </div>
    <div class="right" id="rightdiv">
      <div class="buttons" id="buttonsdiv">
        <button id="prev">Prev</button><button id="next">Next</button>
        <div>
          <span id="npage"></span>
          <span>/</span>
          <span id="npages"></span>
        </div>
        <button id="clr">Clear Page</button><button id="ca">Claer All</button>
      </div>
      <div id="cnvdiv">
        <div class="stack" id="cnvcont">
          <canvas id="cnv"></canvas>
          <canvas id="drw_cnv"></canvas>
        </div>
      </div>
    </div>
  </div>
 </body>
 </html>
--- a/requirements.txt
+++ b/requirements.txt
@@ -44,10 +44,12 @@ pypdf==5.2.0
 pytesseract==0.3.13
 python-dotenv==1.0.1
 python-multipart==0.0.20
 pytz==2025.2
 PyYAML==6.0.2
 requests==2.32.3
 rich==13.9.4
 rich-toolkit==0.13.2
 schedule==1.2.2
 shellingham==1.5.4
 sniffio==1.3.1
 starlette==0.45.3
Author	SHA1	Message	Date
Marcel Gansfusz	6d5c390350	fixed error in compose.yml	2025-11-04 21:24:38 +01:00
Marcel Gansfusz	e13d06d4a7	fixed regular deletions of files	2025-11-04 19:16:59 +01:00
Marcel Gansfusz	547411ba03	made to kill fip entrys when 1d passes	2025-11-04 19:04:41 +01:00
Marcel Gansfusz	cdd26e0bc3	caught exception in trying to censor	2025-11-04 17:54:24 +01:00
Marcel Gansfusz	f7c73a0c5a	fixed js	2025-11-04 17:17:13 +01:00
Marcel Gansfusz	8e74848397	fixed js	2025-11-04 17:13:18 +01:00
Marcel Gansfusz	8704aee82e	fixed tesseract in Dockerfile	2025-11-04 16:56:24 +01:00
Marcel Gansfusz	594ac1fa00	updated censoring status logic	2025-11-04 16:55:11 +01:00
Marcel Gansfusz	2ee90cd0d7	added tessercat to Dockerfile	2025-11-04 15:45:37 +01:00
Marcel Gansfusz	d42bab5b19	changed the fetch in js to be relative (no explicite url; just a path); removed version from docker compose	2025-11-04 14:55:04 +01:00
Marcel Gansfusz	c3a87ceee6	changed stryle of greeting file	2025-10-31 17:48:40 +01:00
Marcel Gansfusz	6f2d373292	updated greeting file to represent new censoring mechanism	2025-10-31 16:18:43 +01:00
Marcel Gansfusz	a37206d6a4	added logging statement	2025-10-30 15:48:51 +01:00
Marcel Gansfusz	6bd75bf93f	removed .nvim; added log statements	2025-10-30 15:31:00 +01:00
Marcel Gansfusz	5bc24a32d5	removed __pycache__	2025-10-30 15:09:16 +01:00
Marcel Gansfusz	a9233926e5	added logging statements	2025-10-30 14:45:53 +01:00
Marcel Gansfusz	90235d2788	Made the database reconnect when connection is broken	2025-10-30 13:03:02 +01:00
Marcel Gansfusz	da316a9351	changed from string paths tp pathlib	2025-10-29 12:14:32 +01:00
Marcel Gansfusz	e6727daf8e	i forgor	2025-10-28 19:32:33 +01:00
Marcel Gansfusz	d6508c739d	in between state before converting to pathlib	2025-10-28 19:32:01 +01:00
Marcel Gansfusz	856c401c06	moved DOCKERFILE to Dockerfile	2025-10-27 18:17:43 +01:00
Marcel Gansfusz	4da77c95d1	finished writeing compatibility with docker; untested	2025-10-24 21:19:36 +02:00
Marcel Gansfusz	98742107b2	changed structure for docker usage	2025-10-24 21:02:42 +02:00
Marcel Gansfusz	b9eb5e8bd4	Merge pull request 'improve_censoring_speed' (#1 ) from improve_censoring_speed into main Reviewed-on: #1	2025-10-23 15:43:40 +02:00
Marcel Gansfusz	5c6a8dfba2	fixed bug in js, that blocked showing prof suggestions when nothing is entered in the field	2025-10-23 15:40:45 +02:00
Marcel Gansfusz	c30d69d205	added back option to run OCR	2025-10-23 00:06:25 +02:00
Marcel Gansfusz	56d3468889	changed the censoring mode to built in censoring with pymupdf	2025-10-22 23:26:33 +02:00
Marcel Gansfusz	26ea274023	made the loading animation prettier	2025-10-22 21:11:40 +02:00
Marcel Gansfusz	0c96d04326	Added Loading circle	2025-10-22 18:10:56 +02:00
Marcel Gansfusz	352540a3b1	Added afunctionality to readout wich page is currently being censored; It probably helps with patience; It is at the moment only implemented in the backend; A pretty frontend is still nessecerry	2025-10-22 01:12:25 +02:00
Marcel Gansfusz	7d828a7c3b	modified the init file to work on local directory; added logging to init.py	2025-10-21 14:48:06 +02:00
Marcel Gansfusz	93f2c59997	prevented overwriteing of files and modified the filename of things that require a date	2025-10-20 15:04:34 +02:00
		`@@ -1,2 +0,0 @@`
			`# remote_path="/srv/http/"`
			`# remote_path="dev@10.0.0.25:/var/www/html/"`