add ocr code

ocr/label_captchas_streamlit.py

@@ -0,0 +1,56 @@
+import streamlit as st
+import os
+import csv
+from PIL import Image
+
+"""Einfaches Streamlit-Tool zum manuellen Labeln von Captcha-Bildern."""
+
+CAPTCHA_DIR = "./captchas"
+CSV_FILE = "captcha_labels.csv"
+
+def load_labeled():
+    if not os.path.exists(CSV_FILE):
+        return set()
+    with open(CSV_FILE, newline='') as f:
+        return set(row[0] for row in csv.reader(f))
+
+def save_label(filename, label):
+    with open(CSV_FILE, "a", newline='') as f:
+        writer = csv.writer(f)
+        writer.writerow([filename, label])
+
+def get_unlabeled_files():
+    labeled = load_labeled()
+    return [f for f in os.listdir(CAPTCHA_DIR) if f.endswith(".png") and f not in labeled]
+
+st.title("Captcha Labeler")
+
+files = get_unlabeled_files()
+if not files:
+    st.success("Alle Captchas sind gelabelt!")
+else:
+    if "idx" not in st.session_state:
+        st.session_state.idx = 0
+    if "flash" not in st.session_state:
+        st.session_state.flash = ""
+    if st.session_state.flash:
+        st.success(st.session_state.flash)
+        st.session_state.flash = ""  # Nach Anzeige zurücksetzen
+
+    if st.button("Nächstes Captcha"):
+        st.session_state.idx += 1
+        st.rerun()
+    if st.session_state.idx >= len(files):
+        st.success("Alle Captchas sind gelabelt!")
+    else:
+        fname = files[st.session_state.idx]
+        img = Image.open(os.path.join(CAPTCHA_DIR, fname))
+        st.image(img, caption=fname)
+        with st.form(key=f"form_{fname}"):
+            label = st.text_input("Lösung eingeben", key=f"label_{fname}")
+            submitted = st.form_submit_button("Speichern")
+            if submitted and label.strip():
+                save_label(fname, label.strip())
+                st.session_state.flash = f"Gespeichert: {fname} -> {label.strip()}"
+                st.session_state.idx += 1
+                st.rerun()