Merge branch 'staging' into production

mathtext_fastapi/nlu.py

@@ -1,9 +1,12 @@
+import re
+
 from collections.abc import Mapping
 from logging import getLogger
 import datetime as dt
 from dateutil.parser import isoparse
 
 from fuzzywuzzy import fuzz
+from fuzzywuzzy import process
 from mathtext_fastapi.intent_classification import predict_message_intent
 from mathtext_fastapi.logging import prepare_message_data_for_logging
 from mathtext.sentiment import sentiment
@@ -121,7 +124,7 @@ def run_intent_classification(message_text):
     label = ''
     ratio = 0
     nlu_response = {'type': 'intent', 'data': label, 'confidence': ratio}
-    commands = [
+    keywords = [
         'easier',
         'exit',
         'harder',
@@ -141,15 +144,21 @@ def run_intent_classification(message_text):
         'skip',
         'menu'
     ]
-    
-    for command in commands:
+
+    try:
+        tokens = re.findall(r"[-a-zA-Z'_]+", message_text.lower())
+    except AttributeError:
+        tokens = ''
+
+    for keyword in keywords:
         try:
-            ratio = fuzz.ratio(command, message_text.lower())
+            tok, score = process.extractOne(keyword, tokens, scorer=fuzz.ratio)
         except:
-            ratio = 0
-        if ratio > 80:
-            nlu_response['data'] = command
-            nlu_response['confidence'] = ratio / 100
+            score = 0
+
+        if score > 80:
+            nlu_response['data'] = keyword
+            nlu_response['confidence'] = score
     
     return nlu_response
 

scripts/make_request.py

@@ -77,11 +77,14 @@ def run_full_nlu_endpoint_payload_test(sample_payload):
 # # Case: Correct payload
 # run_full_nlu_endpoint_payload_test(b'{"message_data": {"author_id": "57787919091", "author_type": "OWNER", "contact_uuid": "df78gsdf78df", "message_body": "8", "message_direction": "inbound", "message_id": "dfgha789789ag9ga", "message_inserted_at": "2023-01-10T02:37:28.487319Z", "message_updated_at": "2023-01-10T02:37:28.487319Z"}}')
 
+# Case: Correct payload + extra fields
+run_full_nlu_endpoint_payload_test(b'{"message_data": {"author_id": "57787919091", "author_type": "OWNER", "contact_uuid": "df78gsdf78df", "message_body": "8", "message_direction": "inbound", "message_id": "dfgha789789ag9ga", "message_inserted_at": "2023-01-10T02:37:28.487319Z", "message_updated_at": "2023-01-10T02:37:28.487319Z", "question": "What is next - 2, 6, 8?", "expected_answer": 8}}')
+
 # # Case: Incorrect payload values
 # run_full_nlu_endpoint_payload_test(b'{"message_data": {"author_id": "@event.message._vnd.v1.chat.owner", "author_type": "@event.message._vnd.v1.author.type", "contact_uuid": "@event.message._vnd.v1.chat.contact_uuid", "message_body": "@event.message.text.body", "message_direction": "@event.message._vnd.v1.direction", "message_id": "@event.message.id", "message_inserted_at": "@event.message._vnd.v1.chat.inserted_at", "message_updated_at": "@event.message._vnd.v1.chat.updated_at"}}')
 
 # Case: Wrong payload object
-run_full_nlu_endpoint_payload_test(b'{"message_data": {"_vnd": {"v1": {"author": {"id": 54327547257, "name": "Jin", "type": "OWNER"}, "card_uuid": None, "chat": {"assigned_to": None, "contact_uuid": "f7889-f78dfgb798-f786ah89g7-f78f9a", "inserted_at": "2023-03-28T13:21:47.581221Z", "owner": "+43789789146", "permalink": "", "state": "OPEN", "state_reason": "Re-opened by inbound message.", "unread_count": 97, "updated_at": "2023-04-07T21:05:27.389948Z", "uuid": "dfg9a78-d76a786dghas-78d9fga789g-a78d69a9"}, "direction": "inbound", "faq_uuid": None, "in_reply_to": None, "inserted_at": "2023-04-07T21:05:27.368580Z", "labels": [], "last_status": None, "last_status_timestamp": None, "on_fallback_channel": False, "rendered_content": None, "uuid": "hf78s7s89b-789fb68d9fg-789fb789dfb-f79sfb789"}}, "from": 5475248689, "id": "SBDE4zgAAy7887sfdT35SHFS", "text": {"body": 1000}, "timestamp": 1680901527, "type": "text"}, "type": "message"}')
+# run_full_nlu_endpoint_payload_test(b'{"message_data": {"_vnd": {"v1": {"author": {"id": 54327547257, "name": "Jin", "type": "OWNER"}, "card_uuid": None, "chat": {"assigned_to": None, "contact_uuid": "f7889-f78dfgb798-f786ah89g7-f78f9a", "inserted_at": "2023-03-28T13:21:47.581221Z", "owner": "+43789789146", "permalink": "", "state": "OPEN", "state_reason": "Re-opened by inbound message.", "unread_count": 97, "updated_at": "2023-04-07T21:05:27.389948Z", "uuid": "dfg9a78-d76a786dghas-78d9fga789g-a78d69a9"}, "direction": "inbound", "faq_uuid": None, "in_reply_to": None, "inserted_at": "2023-04-07T21:05:27.368580Z", "labels": [], "last_status": None, "last_status_timestamp": None, "on_fallback_channel": False, "rendered_content": None, "uuid": "hf78s7s89b-789fb68d9fg-789fb789dfb-f79sfb789"}}, "from": 5475248689, "id": "SBDE4zgAAy7887sfdT35SHFS", "text": {"body": 1000}, "timestamp": 1680901527, "type": "text"}, "type": "message"}')
 
 
 # run_simulated_request('intent-classification', 'exit')