added back option to run OCR

app/main.py

@@ -414,14 +414,14 @@ 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"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)
@@ -454,15 +454,24 @@ async def get_submission(
     censor_status_datas[fileId] = {}
     if fileId not in censor_status_update_events:
         censor_status_update_events[fileId] = asyncio.Event()
-    await asyncio.to_thread(
-        censor_pdf,
-        filepath,
-        dest,
-        rects_p,
-        scales_p,
-        False if censor == "False" else True,
-        fileId,
-    )
+    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)
@@ -505,7 +514,6 @@ def censor_pdf(
     destpath: str,
     rects: List[List[List[float]]],
     scales: List[Dict[str, float]],
-    secure: bool,
     file_id: str,
 ):
     """Censors pdf and saves the file to the given Destpath.
@@ -549,6 +557,65 @@ def censor_pdf(
     censor_status_update_events[file_id].set()
 
 
+def censor_pdf_ocr(
+    path: str,
+    destpath: str,
+    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
+    """
+    doc = pymupdf.open(path)
+    output = pymupdf.open()
+    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.draw_rect(
+                    prect,
+                    color=(0, 0, 0),
+                    fill=(0, 0, 0),
+                )
+        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()
+        # THis Costs us dearly
+        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))
+        # End of the costly part
+        print(f"Page {i + 1}/{npage}: 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:
     return b"\x00\x66\x99"
 

index.html

@@ -134,11 +134,10 @@
             </div>
             <input
               type="checkbox"
-              name="censor"
+              name="ocr"
               id="sec_censor"
               value="True"
-              checked
-            /><label for="sec_censor">Zensieren</label><br /><br />
+            /><label for="sec_censor">OCR</label><br /><br />
             <button type="submit" id="send">Senden</button>
           </form>
         </div>

static/app.js

@@ -276,8 +276,8 @@ function submitPdf(eve) {
   formdata.append("fileId", doc.fID);
   //formdata.append("filename", doc.filename);
   formdata.append("ftype", doc.filetype);
-  if (!formdata.has("censor")) {
-    formdata.append("censor", "False");
+  if (!formdata.has("ocr")) {
+    formdata.append("ocr", "False");
   }
   console.log(formdata);
   submitForm(formdata);