[WIP] Feat/dff/voice skill Oleg

annotators/IntentCatcherTransformers/intent_phrases_commands.json

@@ -1,5 +1,19 @@
 {
     "intent_phrases": {
+        "test_command": {
+            "phrases": [
+                "test_command",
+                "test command"
+            ],
+            "reg_phrases": [
+                "test_command",
+                "test command"
+            ],
+            "min_precision": 0.94,
+            "punctuation": [
+                "."
+            ]
+        },
         "track_object": {
             "phrases": [
                 "((track)|(follow)|(trail)|(trace)|(find)|(rail)|(groove)|(monitor)) a ((human)|(man)|(car)|(bicycle)|(girl)|(dude)|(bag)|(chair)|(black dog)|(white cat))",

annotators/IntentCatcherTransformers/test.sh

@@ -1,3 +1,3 @@
 #!/bin/bash
 
-python test.py
+python tests/test.py

annotators/IntentCatcherTransformers/tests/conftest.py

@@ -0,0 +1,46 @@
+import json
+
+import pytest
+
+from os import getenv
+
+INTENT_PHRASES_PATH = getenv("INTENT_PHRASES_PATH")
+
+
+def pytest_addoption(parser):
+    parser.addoption("--uri", action="store", default="http://0.0.0.0")
+    parser.addoption("--port", action="store", default="8014")
+    parser.addoption("--handle", action="store", default="detect")
+
+
+@pytest.fixture
+def uri(request) -> str:
+    return request.config.getoption("--uri")
+
+
+@pytest.fixture
+def port(request) -> str:
+    return request.config.getoption("--port")
+
+
+@pytest.fixture
+def handle(request) -> str:
+    return request.config.getoption("--handle")
+
+
+@pytest.fixture
+def url(uri, port, handle) -> str:
+    return f"{uri}:{port}/{handle}"
+
+
+@pytest.fixture
+def tests():
+    if "RU" in INTENT_PHRASES_PATH and "commands" in INTENT_PHRASES_PATH:
+        tests = json.load(open("tests_commands_RU.json"))
+    elif "RU" in INTENT_PHRASES_PATH:
+        tests = json.load(open("tests_RU.json"))
+    elif "commands" in INTENT_PHRASES_PATH:
+        tests = json.load(open("tests_commands.json"))
+    else:
+        tests = json.load(open("tests.json"))
+    return tests

annotators/IntentCatcherTransformers/test.py → annotators/IntentCatcherTransformers/tests/test.py

@@ -2,23 +2,9 @@
 
 import requests
 import json
-from os import getenv
 
-INTENT_PHRASES_PATH = getenv("INTENT_PHRASES_PATH")
-SERVICE_PORT = getenv("SERVICE_PORT")
-
-
-def main_test():
-    url = f"http://0.0.0.0:{SERVICE_PORT}/detect"
-    if "RU" in INTENT_PHRASES_PATH and "commands" in INTENT_PHRASES_PATH:
-        tests = json.load(open("tests_commands_RU.json"))
-    elif "RU" in INTENT_PHRASES_PATH:
-        tests = json.load(open("tests_RU.json"))
-    elif "commands" in INTENT_PHRASES_PATH:
-        tests = json.load(open("tests_commands.json"))
-    else:
-        tests = json.load(open("tests.json"))
 
+def test_intent_catcher(url: str, tests: dict):
     for test in tests:
         r = requests.post(url=url, json={"sentences": [[test["sentence"]]]})
         assert r.ok
@@ -30,8 +16,3 @@ def main_test():
             ), print(f"TEST FAILED!\nTest: {test}\nResult:{json.dumps(data, indent=2)}")
         else:
             assert all([intent["detected"] == 0 for intent in data.values()]), f"test: {test}\nprediction: {data}"
-    print("Success")
-
-
-if __name__ == "__main__":
-    main_test()

annotators/IntentCatcherTransformers/utils.py

@@ -3,7 +3,7 @@
 from itertools import chain
 from typing import List
 
-from common.universal_templates import join_sentences_in_or_pattern
+from common.join_pattern import join_sentences_in_or_pattern
 
 
 def get_regexp(intent_phrases_path):

annotators/NER/tests/conftest.py

@@ -0,0 +1,27 @@
+import pytest
+
+
+def pytest_addoption(parser):
+    parser.addoption("--uri", action="store", default="http://proxy.deeppavlov.ai")
+    parser.addoption("--port", action="store", default=8021)
+    parser.addoption("--handle", action="store", default="/ner")
+
+
+@pytest.fixture
+def uri(request) -> str:
+    return request.config.getoption("--uri")
+
+
+@pytest.fixture
+def port(request) -> int:
+    return request.config.getoption("--port")
+
+
+@pytest.fixture
+def handle(request) -> str:
+    return request.config.getoption("--handle")
+
+
+@pytest.fixture
+def url(uri, port, handle) -> str:
+    return f"{uri}:{port}/{handle}"

annotators/NER/tests/test_ner.py

@@ -0,0 +1,41 @@
+import json
+
+import pytest
+
+import sys
+from os import path
+
+import requests
+
+PARENT_DIR = path.dirname(path.dirname(path.abspath(__file__)))
+sys.path.append(PARENT_DIR)
+
+
+@pytest.mark.parametrize(
+    "sentences, gold_result",
+    [
+        (
+            {
+                "last_utterances": [
+                    ["john peterson is my brother.", "he lives in New York."],
+                    ["my laptop was broken.", "could you show me the nearest store in Moscow where i can fix it."],
+                ]
+            },
+            [
+                [
+                    [{"confidence": 1, "end_pos": 2, "start_pos": 0, "text": "john peterson", "type": "PER"}],
+                    [{"confidence": 1, "end_pos": 5, "start_pos": 3, "text": "New York", "type": "LOC"}],
+                ],
+                [
+                    [],
+                    [{"confidence": 1, "end_pos": 9, "start_pos": 8, "text": "Moscow", "type": "LOC"}],
+                ],
+            ],
+        )
+    ],
+)
+def test_ner(url: str, sentences: dict, gold_result: list):
+    response = requests.post(url, json=sentences, headers={"Content-Type": "application/json"})
+    result = json.loads(response.text)
+    assert response.status_code == 200
+    assert result == gold_result

annotators/NER_deeppavlov/Dockerfile-test

@@ -0,0 +1,20 @@
+FROM deeppavlov/deeppavlov:1.2.0-gpu
+
+ARG CONFIG
+ARG SERVICE_PORT
+ARG SRC_DIR
+ARG SED_ARG=" | "
+
+ENV CONFIG=$CONFIG
+ENV SERVICE_PORT=$SERVICE_PORT
+
+COPY ./annotators/NER_deeppavlov/tests/requirements.txt /src/requirements.txt
+
+RUN pip install --upgrade pip && \
+    pip install -r /src/requirements.txt
+
+COPY $SRC_DIR /src
+
+WORKDIR /src
+
+CMD gunicorn  --workers=1 --timeout 800 server:app -b 0.0.0.0:8021

annotators/NER_deeppavlov/test.sh

@@ -1,4 +1,4 @@
 #!/bin/bash
 
 
-python test_server.py
+python -m pytest tests/test_server.py

annotators/NER_deeppavlov/tests/conftest.py

@@ -0,0 +1,27 @@
+import pytest
+
+
+def pytest_addoption(parser):
+    parser.addoption("--uri", action="store", default="http://0.0.0.0")
+    parser.addoption("--port", action="store", default=8021)
+    parser.addoption("--handle", action="store", default="/ner")
+
+
+@pytest.fixture
+def uri(request) -> str:
+    return request.config.getoption("--uri")
+
+
+@pytest.fixture
+def port(request) -> int:
+    return request.config.getoption("--port")
+
+
+@pytest.fixture
+def handle(request) -> str:
+    return request.config.getoption("--handle")
+
+
+@pytest.fixture
+def url(uri, port, handle) -> str:
+    return f"{uri}:{port}/{handle}"

annotators/NER_deeppavlov/tests/requirements.txt

@@ -0,0 +1,11 @@
+sentry-sdk[flask]==0.14.1
+flask==1.1.1
+gunicorn==19.9.0
+itsdangerous==2.0.1
+jinja2<=3.0.3
+Werkzeug<=2.0.3
+transformers==4.6.0
+datasets==1.11.0
+huggingface-hub==0.0.8
+pytest==7.4.2
+allure-pytest==2.13.2

annotators/NER_deeppavlov/tests/test_server.py

@@ -0,0 +1,56 @@
+import allure
+import pytest
+import requests
+
+
+@allure.description("""Test NER""")
+@pytest.mark.parametrize(
+    "request_data, gold_results",
+    [
+        (
+            {
+                "last_utterances": [
+                    ["я видела ивана в москве"],
+                    ["Я видела Ивана в Москве"],
+                    ["i have heard about justin. he is in sahara desert"],
+                    ["I have heard about Justin. He is in Sahara Desert"],
+                    [
+                        "can john smith move forward for 15 meters, then for \
+                    fifteen meters, and get back to las vegas then"
+                    ],
+                    ["я бы проехала на 30 метров вперед, а потом повернула на сорок пять градусов по часовой стрелке"],
+                    [""],
+                ]
+            },
+            [
+                [[]],
+                [[]],
+                [
+                    [
+                        {"start_pos": 4, "end_pos": 5, "type": "PER", "text": "justin", "confidence": 1},
+                        {"start_pos": 9, "end_pos": 11, "type": "LOC", "text": "sahara desert", "confidence": 1},
+                    ]
+                ],
+                [
+                    [
+                        {"start_pos": 4, "end_pos": 5, "type": "PER", "text": "Justin", "confidence": 1},
+                        {"start_pos": 9, "end_pos": 11, "type": "LOC", "text": "Sahara Desert", "confidence": 1},
+                    ]
+                ],
+                [
+                    [
+                        {"start_pos": 1, "end_pos": 3, "type": "PER", "text": "john smith", "confidence": 1},
+                        {"start_pos": 6, "end_pos": 8, "type": "QUANTITY", "text": "15 meters", "confidence": 1},
+                        {"start_pos": 11, "end_pos": 13, "type": "QUANTITY", "text": "fifteen meters", "confidence": 1},
+                        {"start_pos": 18, "end_pos": 20, "type": "LOC", "text": "las vegas", "confidence": 1},
+                    ]
+                ],
+                [[]],
+                [[]],
+            ],
+        )
+    ],
+)
+def test_ner(url: str, request_data: dict, gold_results: list):
+    result = requests.post(url, json=request_data).json()
+    assert result == gold_results

annotators/SentSeg/Dockerfile-test

@@ -0,0 +1,28 @@
+FROM python:3.7-slim
+
+ARG DATA_URL=files.deeppavlov.ai/alexaprize_data/sentseg/elmo2.tar.gz
+ARG MODEL_META_URL=files.deeppavlov.ai/alexaprize_data/sentseg/model.meta
+ARG MODEL_DATA_URL=files.deeppavlov.ai/alexaprize_data/sentseg/model.data-00000-of-00001
+
+WORKDIR /src
+RUN mkdir /data /elmo2 tfhub_cache_dir
+
+RUN apt-get update && \
+    apt-get install -y curl && \
+    curl -L $DATA_URL --output /tmp/elmo2.tar.gz && \
+    tar -xf /tmp/elmo2.tar.gz -C /elmo2 && \
+    rm /tmp/elmo2.tar.gz && \
+    curl -L $MODEL_META_URL --output /data/model.meta && \
+    curl -L $MODEL_DATA_URL --output /data/model.data-00000-of-00001
+
+ENV TFHUB_CACHE_DIR tfhub_cache_dir
+
+COPY tests/requirements.txt .
+RUN pip install --upgrade pip && \
+    pip install -r requirements.txt && \
+    python -c "import nltk; nltk.download('punkt')"
+
+COPY . .
+COPY model.index /data/
+
+CMD gunicorn --workers=1 server:app