- Popravljena swagger definicija in json (/oznaciBesedilo ni delal)

- Dodani resourci za classlo
- Priprava util-ov za podatkovno bazo (se v zelo surovem stanju)
- Dodan docker-compose, ki bo potreben za nadaljni development
- Popravilejni controllerji

swagger_server/controllers/marktext_controller.py

@@ -1,7 +1,9 @@
 import connexion
 import six
 
+from swagger_server.models.oznaci_besedilo_body import OznaciBesediloBody  # noqa: E501
 from swagger_server import util
+from swagger_server.classla import cl_utils
 
 
 def get_text(body):  # noqa: E501
@@ -9,11 +11,13 @@ def get_text(body):  # noqa: E501
 
      # noqa: E501
 
-    :param body: 
+    :param body:
     :type body: dict | bytes
 
     :rtype: str
     """
     if connexion.request.is_json:
-        body = str.from_dict(connexion.request.get_json())  # noqa: E501
-    return 'do some magic!'
+        body = OznaciBesediloBody.from_dict(connexion.request.get_json())  # noqa: E501
+    conllu = cl_utils.raw_text_to_conllu(body.besedilo)
+    return conllu
+

swagger_server/controllers/oss_controller.py

@@ -1,8 +1,11 @@
 import connexion
 import six
+import os
 
 from swagger_server.models.terminoloski_kandidat import TerminoloskiKandidat  # noqa: E501
 from swagger_server import util
+from flask import send_file
+from swagger_server.db_utils import Ngrams_Manager
 
 
 def get_conllu(file_id):  # noqa: E501
@@ -15,7 +18,17 @@ def get_conllu(file_id):  # noqa: E501
 
     :rtype: str
     """
-    return 'do some magic!'
+    file_path = util.get_conllu_path_by_id(file_id)
+
+    a = Ngrams_Manager.get_by_file_id(file_id)
+    print(a)
+
+    try:
+        return send_file(file_path, attachment_filename=f'{file_id}.conllu', as_attachment=True)
+    except Exception as e:
+        if "The system cannot find the file specified" in str(e):
+            return "The conllu with this ID doesn't exist.", 404
+        return str(e), 400
 
 
 def get_conllus(leta, vrste, kljucnebesede, cerifpodrocja):  # noqa: E501

swagger_server/db_utils.py

@@ -0,0 +1,52 @@
+import mariadb
+import sys
+
+
+# todo: DO NOT PUSH THIS TO GIT
+database_info = {
+    'database': 'TEMPORARYVALJUSTFORTHECOMMITPURPOSE',
+    'host': 'TEMPORARYVALJUSTFORTHECOMMITPURPOSE',
+    'port': 'TEMPORARYVALJUSTFORTHECOMMITPURPOSE',
+    'user': 'TEMPORARYVALJUSTFORTHECOMMITPURPOSE',
+    'password': 'TEMPORARYVALJUSTFORTHECOMMITPURPOSE'
+}
+
+# Connect to MariaDB Platform
+try:
+    conn = mariadb.connect(**database_info)
+except mariadb.Error as e:
+    print(f"Error connecting to MariaDB Platform: {e}")
+    sys.exit(1)
+
+# Get Cursor
+cur = conn.cursor()
+
+# class BaseModel(Model):
+#     class Meta:
+#         database = db
+#
+#
+# class os2022_ngrams(BaseModel):
+#     file_id = IntegerField()
+#     sent_id = FloatField()
+#     ngram_len = IntegerField()
+#     frequency_g_t = IntegerField()
+#     gram_text = TextField()
+#     lemma_text = TextField()
+#     xpos_text = TextField()
+#     upos_text = TextField()
+#
+# db.connect()
+
+
+class Ngrams_Manager:
+    @staticmethod
+    def get_by_file_id(file_id):
+        try:
+            # cur.execute(f'SELECT * from os2022_ngrams WHERE file_id = {file_id}')
+            # cur.execute(f'SELECT COUNT(*) FROM os2022_ngrams')
+            # return list(cur)
+            return 1
+        except Exception as e:
+            print(e, 'EXC')
+        return 0

swagger_server/models/__init__.py

@@ -6,4 +6,5 @@ from __future__ import absolute_import
 from swagger_server.models.datoteka_v_besedilo_body import DatotekaVBesediloBody
 from swagger_server.models.datoteka_v_besedilo_ocr_body import DatotekaVBesediloOcrBody
 from swagger_server.models.izlusci_body import IzlusciBody
+from swagger_server.models.oznaci_besedilo_body import OznaciBesediloBody
 from swagger_server.models.terminoloski_kandidat import TerminoloskiKandidat

swagger_server/models/oznaci_besedilo_body.py

@@ -0,0 +1,62 @@
+# coding: utf-8
+
+from __future__ import absolute_import
+from datetime import date, datetime  # noqa: F401
+
+from typing import List, Dict  # noqa: F401
+
+from swagger_server.models.base_model_ import Model
+from swagger_server import util
+
+
+class OznaciBesediloBody(Model):
+    """NOTE: This class is auto generated by the swagger code generator program.
+
+    Do not edit the class manually.
+    """
+    def __init__(self, besedilo: str=None):  # noqa: E501
+        """OznaciBesediloBody - a model defined in Swagger
+
+        :param besedilo: The besedilo of this OznaciBesediloBody.  # noqa: E501
+        :type besedilo: str
+        """
+        self.swagger_types = {
+            'besedilo': str
+        }
+
+        self.attribute_map = {
+            'besedilo': 'besedilo'
+        }
+        self._besedilo = besedilo
+
+    @classmethod
+    def from_dict(cls, dikt) -> 'OznaciBesediloBody':
+        """Returns the dict as a model
+
+        :param dikt: A dict.
+        :type: dict
+        :return: The oznaciBesedilo_body of this OznaciBesediloBody.  # noqa: E501
+        :rtype: OznaciBesediloBody
+        """
+        return util.deserialize_model(dikt, cls)
+
+    @property
+    def besedilo(self) -> str:
+        """Gets the besedilo of this OznaciBesediloBody.
+
+
+        :return: The besedilo of this OznaciBesediloBody.
+        :rtype: str
+        """
+        return self._besedilo
+
+    @besedilo.setter
+    def besedilo(self, besedilo: str):
+        """Sets the besedilo of this OznaciBesediloBody.
+
+
+        :param besedilo: The besedilo of this OznaciBesediloBody.
+        :type besedilo: str
+        """
+
+        self._besedilo = besedilo

swagger_server/swagger/swagger.yaml

@@ -17,7 +17,7 @@ paths:
         content:
           application/json:
             schema:
-              type: string
+              $ref: '#/components/schemas/oznaciBesedilo_body'
         required: true
       responses:
         "200":
@@ -485,6 +485,11 @@ components:
         kanonicnaoblika: kanonicnaoblika
         nosilnautez: 0.8008282
         kandidat: kandidat
+    oznaciBesedilo_body:
+      type: object
+      properties:
+        besedilo:
+          type: string
     izlusci_body:
       type: object
       properties:

swagger_server/test/test_marktext_controller.py

@@ -5,6 +5,7 @@ from __future__ import absolute_import
 from flask import json
 from six import BytesIO
 
+from swagger_server.models.oznaci_besedilo_body import OznaciBesediloBody  # noqa: E501
 from swagger_server.test import BaseTestCase
 
 
@@ -16,7 +17,7 @@ class TestMarktextController(BaseTestCase):
 
         Označi besedilo s classlo/stanzo z uporabo slovenskih modelov ter vrne conll-u format
         """
-        body = 'body_example'
+        body = OznaciBesediloBody()
         response = self.client.open(
             '/oznaciBesedilo',
             method='POST',

swagger_server/util.py

@@ -140,3 +140,7 @@ def _deserialize_dict(data, boxed_type):
     """
     return {k: _deserialize(v, boxed_type)
             for k, v in six.iteritems(data)}
+
+
+def get_conllu_path_by_id(file_id):
+    return f'..\\mnt\\ssd\\ds_ftp\\classla_OS2022\\conll\\rsdo_doc-{file_id}.plainText.conllu'