Dodane backup metode, ce bi slucajno padla tika.

Dockerfile

@@ -3,6 +3,8 @@ FROM python:3.9
 RUN mkdir -p /usr/src/app
 WORKDIR /usr/src/app
 
+RUN apt-get update && apt-get install python3-pil tesseract-ocr libtesseract-dev tesseract-ocr-eng tesseract-ocr-slv tesseract-ocr-script-latn ffmpeg libsm6 libxext6 libgl1 -y
+
 COPY requirements.txt /usr/src/app/
 
 # RUN pip install --upgrade pip

requirements.txt

@@ -11,4 +11,7 @@ pandas==1.3.3
 python-docx==0.8.11
 lxml==4.8.0
 PyPDF2==2.10.4
-requests
+requests
+pytesseract==0.3.10
+opencv-python==4.5.2.54
+numpy==1.20.3

swagger_server/controllers/doc2text_controller.py

@@ -1,6 +1,8 @@
+import os.path
 from tempfile import TemporaryFile
 
 import connexion
+import pytesseract
 import six
 import requests
 import json
@@ -12,13 +14,60 @@ import xml.etree.ElementTree as ET
 from PyPDF2 import PdfReader
 from swagger_server.classla import cl_utils
 import traceback
+import cv2
+import numpy as np
+
+tika_server = "http://tika2:9999/tika_NAROBENURLZANALASC"
 
-tika_server = "http://tika2:9999/tika"
 
 # endpoint below to be used only for development purposes (don't need to run docker)
 # tika_server = "http://rsdo.lhrs.feri.um.si:9998/tika"
 
 
+def extract_text_if_no_tika(file):
+    if file is None:
+        return "No file provided", 400
+    elif "openxmlformats-officedocument.wordprocessingml.document" in file.content_type:
+        content = [p.text for p in docx.Document(file).paragraphs]
+    elif "application/pdf" in file.content_type:
+        reader = PdfReader(file)
+        content = '\n'.join([p.extract_text() for p in reader.pages])
+        content = "ZACASNO UPORABLJEN DRUGI BRALEC KOT TIKA, TA BO SE DODANA KASNEJE...\n\n" + content
+    elif "text/xml" in file.content_type:
+        root = ET.parse(file).getroot()
+        plainText = root.findall('PlainText')
+        if len(plainText) == 0:
+            return "Didn't find anything in PlainText", 400
+        content = '\n'.join([pt.text for pt in plainText])
+    # elif "text/plain" in file.content_type:
+    else:
+        content = file.read().decode('utf-8')
+
+    return content, 200
+
+
+def ocr_text_if_no_tika(file):
+    win_p = "C:/Program Files/Tesseract-OCR/tesseract.exe"
+    if os.path.exists(win_p):
+        pytesseract.pytesseract.tesseract_cmd = win_p
+
+    # convert string data to numpy array
+    file_bytes = np.fromstring(file.read(), np.uint8)
+    # convert numpy array to image
+    img = cv2.imdecode(file_bytes, cv2.IMREAD_COLOR)
+
+    conf = '-l eng+slv'
+    return pytesseract.image_to_string(img, config=conf), 200
+
+
+def tika_responding():
+    try:
+        ret = requests.get(tika_server)
+        return ret.status_code == 200
+    except:
+        return False
+
+
 def datoteka_v_besedilo_post(file=None):  # noqa: E501
     """Pretvori datoteko formata pdf, doc, docx, ppt, xls,... vraca besedilo
 
@@ -32,6 +81,8 @@ def datoteka_v_besedilo_post(file=None):  # noqa: E501
     if file is None:
         return "No file provided", 400
     try:
+        if not tika_responding():
+            return extract_text_if_no_tika(file)
         response = requests.put(tika_server, data=file)
         return response.text, 200
     except Exception as e:
@@ -53,6 +104,8 @@ def get_text_ocr(file=None):  # noqa: E501
     if file is None:
         return "No file provided", 400
     try:
+        if not tika_responding():
+            return ocr_text_if_no_tika(file)
         response = requests.put(tika_server, data=file,
                                 headers={"X-Tika-PDFOcrStrategy": "ocr_only", "X-Tika-OCRLanguage": "slv+eng"})
         return response.text, 200
@@ -75,8 +128,12 @@ def datoteka_v_besedilo_in_classla(file=None):  # noqa: E501
     if file is None:
         return "No file provided", 400
     try:
-        response = requests.put(tika_server, data=file)
+        if tika_responding():
-        return cl_utils.raw_text_to_conllu(response.text)
+            response = requests.put(tika_server, data=file)
+            return cl_utils.raw_text_to_conllu(response.text)
+        else:
+            txt, _ = extract_text_if_no_tika(file)
+            return cl_utils.raw_text_to_conllu(txt)
     except Exception as e:
         return str(e), 500
 
@@ -96,8 +153,12 @@ def get_conllu_ocr(file=None):  # noqa: E501
     if file is None:
         return "No file provided", 400
     try:
-        response = requests.put(tika_server, data=file,
+        if tika_responding():
-                                headers={"X-Tika-PDFOcrStrategy": "ocr_only", "X-Tika-OCRLanguage": "slv+eng"})
+            response = requests.put(tika_server, data=file,
-        return cl_utils.raw_text_to_conllu(response.text)
+                                    headers={"X-Tika-PDFOcrStrategy": "ocr_only", "X-Tika-OCRLanguage": "slv+eng"})
+            return cl_utils.raw_text_to_conllu(response.text)
+        else:
+            txt, _ = ocr_text_if_no_tika(file)
+            return cl_utils.raw_text_to_conllu(txt)
     except Exception as e:
         return str(e), 500