Summarization models

.env

@@ -34,3 +34,4 @@ PROMPT_STORYGPT_SERVICE_URL=http://prompt-storygpt:8127/respond
 STORYGPT_SERVICE_URL=http://storygpt:8126/respond
 SENTENCE_RANKER_SERVICE_URL=http://sentence-ranker:8128/respond
 FILE_SERVER_URL=http://files:3000
+SUMMARIZATION_SERVICE_URL=http://brio-summarizer:8169/respond_batch

.env_ru

@@ -22,3 +22,4 @@ DP_WIKIDATA_URL=http://wiki-parser-ru:8077/model
 DP_ENTITY_LINKING_URL=http://entity-linking-ru:8075/model
 FILE_SERVER_URL=http://files:3000
 SENTENCE_RANKER_SERVICE_URL=http://dialogrpt-ru:8122/rank_sentences
+SUMMARIZATION_SERVICE_URL=http://rut5-summarizer:8170/respond_batch

annotators/summarization_annotator/Dockerfile

@@ -0,0 +1,6 @@
+FROM python:3.7.4
+
+COPY ${WORK_DIR}/requirements.txt /src/requirements.txt
+RUN pip install -r /src/requirements.txt
+COPY ${WORK_DIR} /src
+WORKDIR /src

annotators/summarization_annotator/requirements.txt

@@ -0,0 +1,7 @@
+sentry-sdk[flask]==0.14.1
+flask==1.1.1
+itsdangerous==2.0.1
+gunicorn==19.9.0
+requests==2.22.0
+jinja2<=3.0.3
+Werkzeug<=2.0.3

annotators/summarization_annotator/server.py

@@ -0,0 +1,60 @@
+import logging
+import time
+from os import getenv
+
+import sentry_sdk
+import requests
+from flask import Flask, jsonify, request
+
+
+sentry_sdk.init(getenv("SENTRY_DSN"))
+logging.basicConfig(format="%(asctime)s - %(name)s - %(levelname)s - %(message)s", level=logging.INFO)
+logger = logging.getLogger(__name__)
+app = Flask(__name__)
+
+SUMMARIZATION_SERVICE_URL = getenv("SUMMARIZATION_SERVICE_URL")
+logger.info(f"summarization-annotator considered summarizer: {SUMMARIZATION_SERVICE_URL}")
+
+
+def get_summary(dialog):
+    summary = []
+    if len(dialog) != 11:
+        logger.info(f"summarization-annotator is not ready to summarize dialog as the length of unsummarized dialog is {len(dialog)} != 11")
+        return summary
+
+    logger.info(f"summarization-annotator is ready to summarize dialog as the length of unsummarized dialog is 11")
+    dialog = dialog[:6]
+    for i in range(len(dialog)):
+        if i % 2 == 0:
+            dialog[i] = 'User: ' + dialog[i]
+        else:
+            dialog[i] = 'Bot: ' + dialog[i]
+    dialog = ['\n'.join(dialog)]
+    logger.info(f"summarization-annotator will summarize this: {dialog}")
+
+    try:
+        summary = requests.post(SUMMARIZATION_SERVICE_URL, json={"sentences": dialog}, timeout=10).json()[0]['batch'][0]
+    except Exception as exc:
+        logger.exception(exc)
+        sentry_sdk.capture_exception(exc)
+
+    return summary
+
+
+@app.route("/respond", methods=["POST"])
+def respond():
+    start_time = time.time()
+    dialog = request.json.get('dialog', [])
+
+    logger.info(f"summarization-annotator received dialog: {dialog}")
+    result = get_summary(dialog)
+    summarization_attribute = [{"bot_attributes": {"summarized_dialog": result}}]
+    logger.info(f"summarization-annotator output: {summarization_attribute}")
+
+    total_time = time.time() - start_time
+    logger.info(f"summarization-annotator exec time: {round(total_time, 2)} sec")
+    return jsonify(summarization_attribute)
+
+
+if __name__ == "__main__":
+    app.run(debug=False, host="0.0.0.0", port=8171)

annotators/summarization_annotator/service_configs/summarization-annotator/environment.yml

@@ -0,0 +1,3 @@
+SERVICE_PORT: 8171
+SERVICE_NAME: summarization_annotator
+FLASK_APP: server

annotators/summarization_annotator/service_configs/summarization-annotator/service.yml

@@ -0,0 +1,25 @@
+name: summarization-annotator
+endpoints:
+- respond
+compose:
+  env_file:
+  - .env
+  build:
+    args:
+      SERVICE_PORT: 8171
+      SERVICE_NAME: summarization_annotator
+    context: ./annotators/summarization_annotator/
+  command: flask run -h 0.0.0.0 -p 8171
+  environment:
+  - FLASK_APP=server
+  deploy:
+    resources:
+      limits:
+        memory: 256M
+      reservations:
+        memory: 256M
+  volumes:
+  - ./annotators/summarization_annotator:/src
+  ports:
+  - 8171:8171
+proxy: null

annotators/summarization_annotator/test.py

@@ -0,0 +1,54 @@
+import requests
+from os import getenv
+
+
+SUMMARIZATION_SERVICE_URL = getenv("SUMMARIZATION_SERVICE_URL")
+
+
+def test_skill():
+    url = "http://0.0.0.0:8171/respond"
+
+    if SUMMARIZATION_SERVICE_URL == "http://brio-summarizer:8169/respond_batch":
+        input_data = {"dialog": ["Good morning!",
+                                 "Hi, this is a Dream Socialbot! How is the day going so far for you?",
+                                 "Good! Can you tell me something about cooking and baking?",
+                                 "Sure! Baking cookies is comforting, and cookies are the sweetest "
+                                 "little bit of comfort food. Do you like cooking?",
+                                 "It depends on my mood.",
+                                 "May I recommend you a meal to try to practice cooking?",
+                                 "No. Better tell me what do you have in mind?",
+                                 "I've recently found a couple easy and healthy meals. How about cooking quinoa with "
+                                 "turkey and broccoli?",
+                                 "That sounds like a healthy and tasty meal! Quinoa is a great source of protein, and "
+                                 "when paired with lean turkey and broccoli, it's a well-rounded and balanced meal.",
+                                 "I am glad for you! I listened to my favorite music all day. "
+                                 "Such a great thing you know! Has anything extraordinary happened today?",
+                                 "I can tell you more about what made your day great or we can just chat?"
+                                 "I'm happy to listen!"]}
+
+        desired_output = ["a Dream Socialbot talks to users about cooking and baking cookies. The bot says cookies "
+                          "are comforting, and baking them is a good way to feel good. The robot is called a "
+                          "Dream Social bot. It is designed to talk to users in a friendly, conversational manner."]
+    else:
+        input_data = {"dialog": ["Привет! У тебя есть хобби?",
+                                 "Мое хобби — кулинария.",
+                                 "Здорово! А ты любишь готовить?",
+                                 "Ага, я могу отлично приготовить разные блюда.",
+                                 "Ты собираешь кулинарные рецепты?",
+                                 "Да, уже есть большая коллекция.",
+                                 "А какая национальная кухня тебе нравится?",
+                                 "Конечно, русская.",
+                                 "Русские блюда очень оригинальные, вкусные и полезные.",
+                                 "А что ты любишь готовить больше всего?",
+                                 "Я люблю готовить мясные блюда. Так что приглашаю в гости!"]}
+
+        desired_output = ["У тебя есть хобби — кулинария, а у тебя есть большая коллекция кулинарных рецептов. Bot: Я "
+                          "собираю кулинарные рецепты, собираю кулинарные рецепты, собираю кулинарные рецепты."]
+
+    result = requests.post(url, json=input_data).json()
+    assert result == [{"bot_attributes": {"summarized_dialog": desired_output[0]}}]
+    print("SUCCESS!")
+
+
+if __name__ == "__main__":
+    test_skill()

annotators/summarization_annotator/test.sh

@@ -0,0 +1,3 @@
+#!/bin/bash
+
+python test.py

assistant_dists/dream/dev.yml

@@ -454,6 +454,16 @@ services:
       - "~/.deeppavlov/cache:/root/.cache"
     ports:
       - 8102:8102
+  brio-summarizer:
+    volumes:
+      - "./services/brio_summarizer:/src"
+    ports:
+      - 8169:8169
+  summarization-annotator:
+    volumes:
+      - "./annotators/summarization_annotator:/src"
+    ports:
+      - 8171:8171
   dff-template-skill:
     volumes:
       - "./skills/dff_template_skill:/src"

assistant_dists/dream/docker-compose.override.yml

@@ -1390,6 +1390,42 @@ services:
         reservations:
           memory: 4G
 
+  brio-summarizer:
+    env_file: [ .env ]
+    build:
+      args:
+        SERVICE_PORT: 8169
+        SERVICE_NAME: brio_summarizer
+        PRETRAINED_MODEL_NAME: "Yale-LILY/brio-cnndm-uncased"
+      context: ./services/brio_summarizer/
+    command: flask run -h 0.0.0.0 -p 8169
+    environment:
+      - CUDA_VISIBLE_DEVICES=0
+      - FLASK_APP=server
+    deploy:
+      resources:
+        limits:
+          memory: 4G
+        reservations:
+          memory: 4G
+
+  summarization-annotator:
+    env_file: [ .env ]
+    build:
+      args:
+        SERVICE_PORT: 8171
+        SERVICE_NAME: summarization_annotator
+      context: ./annotators/summarization_annotator/
+    command: flask run -h 0.0.0.0 -p 8171
+    environment:
+      - FLASK_APP=server
+    deploy:
+      resources:
+        limits:
+          memory: 256M
+        reservations:
+          memory: 256M
+
   dff-template-skill:
     env_file: [ .env ]
     build:

assistant_dists/dream/pipeline_conf.json

@@ -515,6 +515,25 @@
                     "component": "components/VkkvnRwjgB5GwrH98k5EKA.yml",
                     "service": "annotators/relative_persona_extractor/service_configs/relative-persona-extractor"
                 }
+            },
+            "summarization_annotator": {
+                "connector": {
+                    "protocol": "http",
+                    "timeout": 10.0,
+                    "url": "http://summarization-annotator:8171/respond"
+                },
+                "dialog_formatter": "state_formatters.dp_formatters:summarization_annotator_formatter",
+                "response_formatter": "state_formatters.dp_formatters:simple_formatter_service",
+                "previous_services": [
+                    "annotators.spelling_preprocessing"
+                ],
+                "state_manager_method": "update_attributes",
+                "is_enabled": true,
+                "source": {
+                    "directory": "annotators/summarization_annotator",
+                    "container": "summarization-annotator",
+                    "endpoint": "respond"
+                }
             }
         },
         "response_annotators": {

assistant_dists/dream_russian/dev.yml

@@ -137,4 +137,14 @@ services:
       - "~/.deeppavlov:/root/.deeppavlov"
     ports:
       - 8078:8078
+  rut5-summarizer:
+    volumes:
+      - "./services/ruT5_summarizer:/src"
+    ports:
+      - 8170:8170
+  summarization-annotator:
+    volumes:
+      - "./annotators/summarization_annotator:/src"
+    ports:
+      - 8171:8171
 version: "3.7"

assistant_dists/dream_russian/docker-compose.override.yml

@@ -445,4 +445,40 @@ services:
         reservations:
           memory: 3G
 
+  rut5-summarizer:
+    env_file: [ .env ]
+    build:
+      args:
+        SERVICE_PORT: 8170
+        SERVICE_NAME: ruT5_summarizer
+        PRETRAINED_MODEL_NAME: "IlyaGusev/rut5_base_sum_gazeta"
+      context: ./services/ruT5_summarizer/
+    command: flask run -h 0.0.0.0 -p 8170
+    environment:
+      - CUDA_VISIBLE_DEVICES=0
+      - FLASK_APP=server
+    deploy:
+      resources:
+        limits:
+          memory: 4G
+        reservations:
+          memory: 4G
+
+  summarization-annotator:
+    env_file: [ .env ]
+    build:
+      args:
+        SERVICE_PORT: 8171
+        SERVICE_NAME: summarization_annotator
+      context: ./annotators/summarization_annotator/
+    command: flask run -h 0.0.0.0 -p 8171
+    environment:
+      - FLASK_APP=server
+    deploy:
+      resources:
+        limits:
+          memory: 256M
+        reservations:
+          memory: 256M
+
 version: '3.7'

assistant_dists/dream_russian/pipeline_conf.json

@@ -266,6 +266,25 @@
                     "component": "components/feDgqHKLibnMNM3HSbnmA.yml",
                     "service": "annotators/wiki_parser/service_configs/wiki-parser-ru"
                 }
+            },
+            "summarization_annotator": {
+                "connector": {
+                    "protocol": "http",
+                    "timeout": 10.0,
+                    "url": "http://summarization-annotator:8171/respond"
+                },
+                "dialog_formatter": "state_formatters.dp_formatters:summarization_annotator_formatter",
+                "response_formatter": "state_formatters.dp_formatters:simple_formatter_service",
+                "previous_services": [
+                    "annotators.spelling_preprocessing"
+                ],
+                "state_manager_method": "update_attributes",
+                "is_enabled": true,
+                "source": {
+                    "directory": "annotators/summarization_annotator",
+                    "container": "summarization-annotator",
+                    "endpoint": "respond"
+                }
             }
         },
         "response_annotators": {

components.tsv

@@ -170,3 +170,6 @@
 8166
 8167	openai-api-chatgpt-16k
 8168	transformers-lm-vicuna13b
+8169	brio-summarizer
+8170	rut5-summarizer
+8171	summarization-annotator

components/riRfdGz86P51B9bL7fO6JR.yml

@@ -0,0 +1,24 @@
+name: summarization-annotator
+display_name: Summarization Annotator
+container_name: summarization-annotator
+component_type: null
+model_type: NN-based
+is_customizable: false
+author: DeepPavlov
+description: Annotator that accesses summarization services
+ram_usage: 256M
+gpu_usage: null
+connector:
+  protocol: http
+  timeout: 10.0
+  url: http://summarization-annotator:8171/respond
+dialog_formatter: state_formatters.dp_formatters:summarization_annotator_formatter
+response_formatter: state_formatters.dp_formatters:simple_formatter_service
+previous_services:
+- annotators.spelling_preprocessing
+required_previous_services: null
+state_manager_method: add_annotation
+tags: null
+endpoint: respond
+service: annotators/summarization_annotator/service_configs/aummarization-annotator
+date_created: '2023-07-04T11:39:32'

services/brio_summarizer/Dockerfile

@@ -0,0 +1,9 @@
+FROM python:3.7.4
+
+ARG PRETRAINED_MODEL_NAME
+ENV PRETRAINED_MODEL_NAME ${PRETRAINED_MODEL_NAME}
+
+COPY ${WORK_DIR}/requirements.txt /src/requirements.txt
+RUN pip install -r /src/requirements.txt
+COPY ${WORK_DIR} /src
+WORKDIR /src

services/brio_summarizer/requirements.txt

@@ -0,0 +1,9 @@
+torch==1.13.1
+transformers==4.27.0
+sentry-sdk[flask]==0.14.1
+flask==1.1.1
+itsdangerous==2.0.1
+gunicorn==19.9.0
+requests==2.22.0
+jinja2<=3.0.3
+Werkzeug<=2.0.3