Annotation management (#1767)

Co-authored-by: jyong <jyong@dify.ai>

api/tasks/annotation/add_annotation_to_index_task.py

@@ -0,0 +1,59 @@
+import logging
+import time
+
+import click
+from celery import shared_task
+from langchain.schema import Document
+
+from core.index.index import IndexBuilder
+
+from models.dataset import Dataset
+from services.dataset_service import DatasetCollectionBindingService
+
+
+@shared_task(queue='dataset')
+def add_annotation_to_index_task(annotation_id: str, question: str, tenant_id: str, app_id: str,
+                                 collection_binding_id: str):
+    """
+    Add annotation to index.
+    :param annotation_id: annotation id
+    :param question: question
+    :param tenant_id: tenant id
+    :param app_id: app id
+    :param collection_binding_id: embedding binding id
+
+    Usage: clean_dataset_task.delay(dataset_id, tenant_id, indexing_technique, index_struct)
+    """
+    logging.info(click.style('Start build index for annotation: {}'.format(annotation_id), fg='green'))
+    start_at = time.perf_counter()
+
+    try:
+        dataset_collection_binding = DatasetCollectionBindingService.get_dataset_collection_binding_by_id_and_type(
+            collection_binding_id,
+            'annotation'
+        )
+        dataset = Dataset(
+            id=app_id,
+            tenant_id=tenant_id,
+            indexing_technique='high_quality',
+            collection_binding_id=dataset_collection_binding.id
+        )
+
+        document = Document(
+            page_content=question,
+            metadata={
+                "annotation_id": annotation_id,
+                "app_id": app_id,
+                "doc_id": annotation_id
+            }
+        )
+        index = IndexBuilder.get_index(dataset, 'high_quality')
+        if index:
+            index.add_texts([document])
+        end_at = time.perf_counter()
+        logging.info(
+            click.style(
+                'Build index successful for annotation: {} latency: {}'.format(annotation_id, end_at - start_at),
+                fg='green'))
+    except Exception:
+        logging.exception("Build index for annotation failed")

api/tasks/annotation/batch_import_annotations_task.py

@@ -0,0 +1,99 @@
+import json
+import logging
+import time
+
+import click
+from celery import shared_task
+from langchain.schema import Document
+from werkzeug.exceptions import NotFound
+
+from core.index.index import IndexBuilder
+from extensions.ext_database import db
+from extensions.ext_redis import redis_client
+from models.dataset import Dataset
+from models.model import MessageAnnotation, App, AppAnnotationSetting
+from services.dataset_service import DatasetCollectionBindingService
+
+
+@shared_task(queue='dataset')
+def batch_import_annotations_task(job_id: str, content_list: list[dict], app_id: str, tenant_id: str,
+                                  user_id: str):
+    """
+    Add annotation to index.
+    :param job_id: job_id
+    :param content_list: content list
+    :param tenant_id: tenant id
+    :param app_id: app id
+    :param user_id: user_id
+
+    """
+    logging.info(click.style('Start batch import annotation: {}'.format(job_id), fg='green'))
+    start_at = time.perf_counter()
+    indexing_cache_key = 'app_annotation_batch_import_{}'.format(str(job_id))
+    # get app info
+    app = db.session.query(App).filter(
+        App.id == app_id,
+        App.tenant_id == tenant_id,
+        App.status == 'normal'
+    ).first()
+
+    if app:
+        try:
+            documents = []
+            for content in content_list:
+                annotation = MessageAnnotation(
+                    app_id=app.id,
+                    content=content['answer'],
+                    question=content['question'],
+                    account_id=user_id
+                )
+                db.session.add(annotation)
+                db.session.flush()
+
+                document = Document(
+                    page_content=content['question'],
+                    metadata={
+                        "annotation_id": annotation.id,
+                        "app_id": app_id,
+                        "doc_id": annotation.id
+                    }
+                )
+                documents.append(document)
+            # if annotation reply is enabled , batch add annotations' index
+            app_annotation_setting = db.session.query(AppAnnotationSetting).filter(
+                AppAnnotationSetting.app_id == app_id
+            ).first()
+
+            if app_annotation_setting:
+                dataset_collection_binding = DatasetCollectionBindingService.get_dataset_collection_binding_by_id_and_type(
+                    app_annotation_setting.collection_binding_id,
+                    'annotation'
+                )
+                if not dataset_collection_binding:
+                    raise NotFound("App annotation setting not found")
+                dataset = Dataset(
+                    id=app_id,
+                    tenant_id=tenant_id,
+                    indexing_technique='high_quality',
+                    embedding_model_provider=dataset_collection_binding.provider_name,
+                    embedding_model=dataset_collection_binding.model_name,
+                    collection_binding_id=dataset_collection_binding.id
+                )
+
+                index = IndexBuilder.get_index(dataset, 'high_quality')
+                if index:
+                    index.add_texts(documents)
+
+            db.session.commit()
+            redis_client.setex(indexing_cache_key, 600, 'completed')
+            end_at = time.perf_counter()
+            logging.info(
+                click.style(
+                    'Build index successful for batch import annotation: {} latency: {}'.format(job_id, end_at - start_at),
+                    fg='green'))
+        except Exception as e:
+            db.session.rollback()
+            redis_client.setex(indexing_cache_key, 600, 'error')
+            indexing_error_msg_key = 'app_annotation_batch_import_error_msg_{}'.format(str(job_id))
+            redis_client.setex(indexing_error_msg_key, 600, str(e))
+            logging.exception("Build index for batch import annotations failed")

api/tasks/annotation/delete_annotation_index_task.py

@@ -0,0 +1,45 @@
+import datetime
+import logging
+import time
+
+import click
+from celery import shared_task
+from core.index.index import IndexBuilder
+from models.dataset import Dataset
+from services.dataset_service import DatasetCollectionBindingService
+
+
+@shared_task(queue='dataset')
+def delete_annotation_index_task(annotation_id: str, app_id: str, tenant_id: str,
+                                 collection_binding_id: str):
+    """
+    Async delete annotation index task
+    """
+    logging.info(click.style('Start delete app annotation index: {}'.format(app_id), fg='green'))
+    start_at = time.perf_counter()
+    try:
+        dataset_collection_binding = DatasetCollectionBindingService.get_dataset_collection_binding_by_id_and_type(
+            collection_binding_id,
+            'annotation'
+        )
+
+        dataset = Dataset(
+            id=app_id,
+            tenant_id=tenant_id,
+            indexing_technique='high_quality',
+            collection_binding_id=dataset_collection_binding.id
+        )
+
+        vector_index = IndexBuilder.get_default_high_quality_index(dataset)
+        if vector_index:
+            try:
+                vector_index.delete_by_metadata_field('annotation_id', annotation_id)
+            except Exception:
+                logging.exception("Delete annotation index failed when annotation deleted.")
+        end_at = time.perf_counter()
+        logging.info(
+            click.style('App annotations index deleted : {} latency: {}'.format(app_id, end_at - start_at),
+                        fg='green'))
+    except Exception as e:
+        logging.exception("Annotation deleted index failed:{}".format(str(e)))
+

api/tasks/annotation/disable_annotation_reply_task.py

@@ -0,0 +1,74 @@
+import datetime
+import logging
+import time
+
+import click
+from celery import shared_task
+from werkzeug.exceptions import NotFound
+
+from core.index.index import IndexBuilder
+from extensions.ext_database import db
+from extensions.ext_redis import redis_client
+from models.dataset import Dataset
+from models.model import MessageAnnotation, App, AppAnnotationSetting
+
+
+@shared_task(queue='dataset')
+def disable_annotation_reply_task(job_id: str, app_id: str, tenant_id: str):
+    """
+    Async enable annotation reply task
+    """
+    logging.info(click.style('Start delete app annotations index: {}'.format(app_id), fg='green'))
+    start_at = time.perf_counter()
+    # get app info
+    app = db.session.query(App).filter(
+        App.id == app_id,
+        App.tenant_id == tenant_id,
+        App.status == 'normal'
+    ).first()
+
+    if not app:
+        raise NotFound("App not found")
+
+    app_annotation_setting = db.session.query(AppAnnotationSetting).filter(
+        AppAnnotationSetting.app_id == app_id
+    ).first()
+
+    if not app_annotation_setting:
+        raise NotFound("App annotation setting not found")
+
+    disable_app_annotation_key = 'disable_app_annotation_{}'.format(str(app_id))
+    disable_app_annotation_job_key = 'disable_app_annotation_job_{}'.format(str(job_id))
+
+    try:
+
+        dataset = Dataset(
+            id=app_id,
+            tenant_id=tenant_id,
+            indexing_technique='high_quality',
+            collection_binding_id=app_annotation_setting.collection_binding_id
+        )
+
+        vector_index = IndexBuilder.get_default_high_quality_index(dataset)
+        if vector_index:
+            try:
+                vector_index.delete_by_metadata_field('app_id', app_id)
+            except Exception:
+                logging.exception("Delete doc index failed when dataset deleted.")
+        redis_client.setex(disable_app_annotation_job_key, 600, 'completed')
+
+        # delete annotation setting
+        db.session.delete(app_annotation_setting)
+        db.session.commit()
+
+        end_at = time.perf_counter()
+        logging.info(
+            click.style('App annotations index deleted : {} latency: {}'.format(app_id, end_at - start_at),
+                        fg='green'))
+    except Exception as e:
+        logging.exception("Annotation batch deleted index failed:{}".format(str(e)))
+        redis_client.setex(disable_app_annotation_job_key, 600, 'error')
+        disable_app_annotation_error_key = 'disable_app_annotation_error_{}'.format(str(job_id))
+        redis_client.setex(disable_app_annotation_error_key, 600, str(e))
+    finally:
+        redis_client.delete(disable_app_annotation_key)

api/tasks/annotation/enable_annotation_reply_task.py

@@ -0,0 +1,106 @@
+import datetime
+import logging
+import time
+
+import click
+from celery import shared_task
+from langchain.schema import Document
+from werkzeug.exceptions import NotFound
+
+from core.index.index import IndexBuilder
+from extensions.ext_database import db
+from extensions.ext_redis import redis_client
+from models.dataset import Dataset
+from models.model import MessageAnnotation, App, AppAnnotationSetting
+from services.dataset_service import DatasetCollectionBindingService
+
+
+@shared_task(queue='dataset')
+def enable_annotation_reply_task(job_id: str, app_id: str, user_id: str, tenant_id: str, score_threshold: float,
+                                 embedding_provider_name: str, embedding_model_name: str):
+    """
+    Async enable annotation reply task
+    """
+    logging.info(click.style('Start add app annotation to index: {}'.format(app_id), fg='green'))
+    start_at = time.perf_counter()
+    # get app info
+    app = db.session.query(App).filter(
+        App.id == app_id,
+        App.tenant_id == tenant_id,
+        App.status == 'normal'
+    ).first()
+
+    if not app:
+        raise NotFound("App not found")
+
+    annotations = db.session.query(MessageAnnotation).filter(MessageAnnotation.app_id == app_id).all()
+    enable_app_annotation_key = 'enable_app_annotation_{}'.format(str(app_id))
+    enable_app_annotation_job_key = 'enable_app_annotation_job_{}'.format(str(job_id))
+
+    try:
+        documents = []
+        dataset_collection_binding = DatasetCollectionBindingService.get_dataset_collection_binding(
+            embedding_provider_name,
+            embedding_model_name,
+            'annotation'
+        )
+        annotation_setting = db.session.query(AppAnnotationSetting).filter(
+            AppAnnotationSetting.app_id == app_id).first()
+        if annotation_setting:
+            annotation_setting.score_threshold = score_threshold
+            annotation_setting.collection_binding_id = dataset_collection_binding.id
+            annotation_setting.updated_user_id = user_id
+            annotation_setting.updated_at = datetime.datetime.utcnow()
+            db.session.add(annotation_setting)
+        else:
+            new_app_annotation_setting = AppAnnotationSetting(
+                app_id=app_id,
+                score_threshold=score_threshold,
+                collection_binding_id=dataset_collection_binding.id,
+                created_user_id=user_id,
+                updated_user_id=user_id
+            )
+            db.session.add(new_app_annotation_setting)
+
+        dataset = Dataset(
+            id=app_id,
+            tenant_id=tenant_id,
+            indexing_technique='high_quality',
+            embedding_model_provider=embedding_provider_name,
+            embedding_model=embedding_model_name,
+            collection_binding_id=dataset_collection_binding.id
+        )
+        if annotations:
+            for annotation in annotations:
+                document = Document(
+                    page_content=annotation.question,
+                    metadata={
+                        "annotation_id": annotation.id,
+                        "app_id": app_id,
+                        "doc_id": annotation.id
+                    }
+                )
+                documents.append(document)
+            index = IndexBuilder.get_index(dataset, 'high_quality')
+            if index:
+                try:
+                    index.delete_by_metadata_field('app_id', app_id)
+                except Exception as e:
+                    logging.info(
+                        click.style('Delete annotation index error: {}'.format(str(e)),
+                                    fg='red'))
+                index.add_texts(documents)
+        db.session.commit()
+        redis_client.setex(enable_app_annotation_job_key, 600, 'completed')
+        end_at = time.perf_counter()
+        logging.info(
+            click.style('App annotations added to index: {} latency: {}'.format(app_id, end_at - start_at),
+                        fg='green'))
+    except Exception as e:
+        logging.exception("Annotation batch created index failed:{}".format(str(e)))
+        redis_client.setex(enable_app_annotation_job_key, 600, 'error')
+        enable_app_annotation_error_key = 'enable_app_annotation_error_{}'.format(str(job_id))
+        redis_client.setex(enable_app_annotation_error_key, 600, str(e))
+        db.session.rollback()
+    finally:
+        redis_client.delete(enable_app_annotation_key)

api/tasks/annotation/update_annotation_to_index_task.py

@@ -0,0 +1,63 @@
+import logging
+import time
+
+import click
+from celery import shared_task
+from langchain.schema import Document
+
+from core.index.index import IndexBuilder
+
+from models.dataset import Dataset
+from services.dataset_service import DatasetCollectionBindingService
+
+
+@shared_task(queue='dataset')
+def update_annotation_to_index_task(annotation_id: str, question: str, tenant_id: str, app_id: str,
+                                    collection_binding_id: str):
+    """
+    Update annotation to index.
+    :param annotation_id: annotation id
+    :param question: question
+    :param tenant_id: tenant id
+    :param app_id: app id
+    :param collection_binding_id: embedding binding id
+
+    Usage: clean_dataset_task.delay(dataset_id, tenant_id, indexing_technique, index_struct)
+    """
+    logging.info(click.style('Start update index for annotation: {}'.format(annotation_id), fg='green'))
+    start_at = time.perf_counter()
+
+    try:
+        dataset_collection_binding = DatasetCollectionBindingService.get_dataset_collection_binding_by_id_and_type(
+            collection_binding_id,
+            'annotation'
+        )
+
+        dataset = Dataset(
+            id=app_id,
+            tenant_id=tenant_id,
+            indexing_technique='high_quality',
+            embedding_model_provider=dataset_collection_binding.provider_name,
+            embedding_model=dataset_collection_binding.model_name,
+            collection_binding_id=dataset_collection_binding.id
+        )
+
+        document = Document(
+            page_content=question,
+            metadata={
+                "annotation_id": annotation_id,
+                "app_id": app_id,
+                "doc_id": annotation_id
+            }
+        )
+        index = IndexBuilder.get_index(dataset, 'high_quality')
+        if index:
+            index.delete_by_metadata_field('annotation_id', annotation_id)
+            index.add_texts([document])
+        end_at = time.perf_counter()
+        logging.info(
+            click.style(
+                'Build index successful for annotation: {} latency: {}'.format(annotation_id, end_at - start_at),
+                fg='green'))
+    except Exception:
+        logging.exception("Build index for annotation failed")