Dodana endpointa za sync pridobivanje connlu-jev, ali preprosto file -> text, ali file -> ocr -> text.

swagger_server/controllers/doc2text_controller.py

@@ -2,7 +2,6 @@ from tempfile import TemporaryFile
 
 import connexion
 import six
-# import tika
 import requests
 import json
 from swagger_server import util
@@ -11,19 +10,20 @@ import docx
 import codecs
 import xml.etree.ElementTree as ET
 from PyPDF2 import PdfReader
+from swagger_server.classla import cl_utils
 
 tika_server = "http://tika:9998/tika"
 
-
-tika_server = "http://rsdo.lhrs.feri.um.si:9998/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 datoteka_v_besedilo_post(file=None):  # noqa: E501
-    """Pretvori datoteko formata pdf, doc, docx, ppt, xls,... v besedilo
+    """Pretvori datoteko formata pdf, doc, docx, ppt, xls,... vraca besedilo
 
      # noqa: E501
 
-    :param file: 
+    :param file:
     :type file: strstr
 
     :rtype: str
@@ -37,10 +37,9 @@ def datoteka_v_besedilo_post(file=None):  # noqa: E501
         return str(e), 500
 
 
-def get_text_ocr(file=None):
+def get_text_ocr(file=None):  # noqa: E501
     """Pretvori datoteko formata pdf, doc, docx, ppt, xls,... v besedilo s pomočjo ocr razpoznavanja
 
-
      # noqa: E501
 
     :param file:
@@ -48,13 +47,54 @@ def get_text_ocr(file=None):
 
     :rtype: str
     """
-
     if file is None:
         return "No file provided", 400
     try:
         response = requests.put(tika_server, data=file,
-                                 headers={"X-Tika-PDFOcrStrategy": "ocr_only", "X-Tika-OCRLanguage": "slv+eng"})
+                                headers={"X-Tika-PDFOcrStrategy": "ocr_only", "X-Tika-OCRLanguage": "slv+eng"})
         return response.text, 200
     except Exception as e:
         return str(e), 500
 
+
+def datoteka_v_besedilo_in_classla(file=None):  # noqa: E501
+    """Pretvori datoteko formata pdf, doc, docx, ppt, xls,... vraca conllu
+
+     # noqa: E501
+
+    :param file:
+    :type file: strstr
+
+    :rtype: str
+    """
+    if not cl_utils.nlp_loaded:
+        return "NLP Models still loading up since server restart, please retry later.", 500
+    if file is None:
+        return "No file provided", 400
+    try:
+        response = requests.put(tika_server, data=file)
+        return cl_utils.raw_text_to_conllu(response.text)
+    except Exception as e:
+        return str(e), 500
+
+
+def get_conllu_ocr(file=None):  # noqa: E501
+    """Pretvori datoteko formata pdf, doc, docx, ppt, xls,... v conllu s pomočjo ocr razpoznavanja
+
+     # noqa: E501
+
+    :param file:
+    :type file: strstr
+
+    :rtype: str
+    """
+    if not cl_utils.nlp_loaded:
+        return "NLP Models still loading up since server restart, please retry later.", 500
+    if file is None:
+        return "No file provided", 400
+    try:
+        response = requests.put(tika_server, data=file,
+                                headers={"X-Tika-PDFOcrStrategy": "ocr_only", "X-Tika-OCRLanguage": "slv+eng"})
+        return cl_utils.raw_text_to_conllu(response.text)
+    except Exception as e:
+        return str(e), 500