From 59cc55b3d96c2ff549f3bdc5f8af1b05f79a715f Mon Sep 17 00:00:00 2001
From: Xu Han <nciiproto@gmail.com>
Date: Fri, 3 Jan 2025 07:27:46 +0000
Subject: [PATCH 1/6] fix: fix intention bedrock issue
---
source/lambda/intention/intention.py | 265 +++++++++++++++------------
1 file changed, 147 insertions(+), 118 deletions(-)
diff --git a/source/lambda/intention/intention.py b/source/lambda/intention/intention.py
index d73213ad..8776712f 100644
--- a/source/lambda/intention/intention.py
+++ b/source/lambda/intention/intention.py
@@ -1,36 +1,43 @@
import hashlib
import json
+import logging
import os
import re
import time
+from io import BytesIO
from typing import List
+
import boto3
-from openpyxl import load_workbook
-from io import BytesIO
-from embeddings import get_embedding_info
+from aos import sm_utils
+from aos.aos_utils import LLMBotOpenSearchClient
from botocore.paginate import TokenEncoder
-from opensearchpy import NotFoundError, RequestError, helpers, RequestsHttpConnection
-import logging
-from langchain.embeddings.bedrock import BedrockEmbeddings
+from constant import (
+ BULK_SIZE,
+ DEFAULT_CONTENT_TYPE,
+ DEFAULT_MAX_ITEMS,
+ DEFAULT_SIZE,
+ DOWNLOAD_RESOURCE,
+ EXECUTION_RESOURCE,
+ INDEX_USED_SCAN_RESOURCE,
+ PRESIGNED_URL_RESOURCE,
+ ModelDimensionMap,
+)
+from embeddings import get_embedding_info
from langchain.docstore.document import Document
+from langchain.embeddings.bedrock import BedrockEmbeddings
from langchain_community.vectorstores import OpenSearchVectorSearch
from langchain_community.vectorstores.opensearch_vector_search import (
OpenSearchVectorSearch,
)
-from aos import sm_utils
+from openpyxl import load_workbook
+from opensearchpy import (
+ NotFoundError,
+ RequestError,
+ RequestsHttpConnection,
+ helpers,
+)
from requests_aws4auth import AWS4Auth
-from aos.aos_utils import LLMBotOpenSearchClient
-from constant import (BULK_SIZE,
- DEFAULT_CONTENT_TYPE,
- DEFAULT_MAX_ITEMS,
- DEFAULT_SIZE,
- DOWNLOAD_RESOURCE,
- EXECUTION_RESOURCE,
- INDEX_USED_SCAN_RESOURCE,
- PRESIGNED_URL_RESOURCE,
- ModelDimensionMap)
-
logger = logging.getLogger(__name__)
encoder = TokenEncoder()
@@ -58,30 +65,36 @@
built_in_tools = ["chat", "get_weather"]
try:
- master_user = sm_client.get_secret_value(
- SecretId=aos_secret)["SecretString"]
+ master_user = sm_client.get_secret_value(SecretId=aos_secret)[
+ "SecretString"
+ ]
secret = json.loads(master_user)
username = secret.get("username")
password = secret.get("password")
if not aos_endpoint:
opensearch_client = boto3.client("opensearch")
- response = opensearch_client.describe_domain(
- DomainName=aos_domain_name)
+ response = opensearch_client.describe_domain(DomainName=aos_domain_name)
aos_endpoint = response["DomainStatus"]["Endpoint"]
- aos_client = LLMBotOpenSearchClient(aos_endpoint, (username, password)).client
+ aos_client = LLMBotOpenSearchClient(
+ aos_endpoint, (username, password)
+ ).client
awsauth = (username, password)
except sm_client.exceptions.ResourceNotFoundException:
logger.info("Secret '%s' not found in Secrets Manager", aos_secret)
aos_client = LLMBotOpenSearchClient(aos_endpoint).client
- awsauth = AWS4Auth(refreshable_credentials=credentials,
- region=region, service="es")
+ awsauth = AWS4Auth(
+ refreshable_credentials=credentials, region=region, service="es"
+ )
except sm_client.exceptions.InvalidRequestException:
- logger.info("InvalidRequestException. It might caused by getting secret value from a deleting secret")
+ logger.info(
+ "InvalidRequestException. It might caused by getting secret value from a deleting secret"
+ )
logger.info("Fallback to authentication with IAM")
aos_client = LLMBotOpenSearchClient(aos_endpoint).client
- awsauth = AWS4Auth(refreshable_credentials=credentials,
- region=region, service="es")
+ awsauth = AWS4Auth(
+ refreshable_credentials=credentials, region=region, service="es"
+ )
except Exception as err:
logger.error("Error retrieving secret '%s': %s", aos_secret, str(err))
raise
@@ -123,7 +136,9 @@ def aos_ingestion(self, documents: List[Document], index: str) -> None:
embeddings_vectors_list.append(
embeddings_vectors[0]["dense_vecs"][doc_id]
)
- metadata["embedding_endpoint_name"] = self.embedding_model_endpoint
+ metadata["embedding_endpoint_name"] = (
+ self.embedding_model_endpoint
+ )
metadata_list.append(metadata)
embeddings_vectors = embeddings_vectors_list
metadatas = metadata_list
@@ -152,10 +167,11 @@ def lambda_handler(event, context):
if resource == PRESIGNED_URL_RESOURCE:
input_body = json.loads(event["body"])
file_name = f"intentions/{group_name}/[{input_body['timestamp']}]{input_body['file_name']}"
- presigned_url = __gen_presigned_url(file_name,
- input_body.get(
- "content_type", DEFAULT_CONTENT_TYPE),
- input_body.get("expiration", 60*60))
+ presigned_url = __gen_presigned_url(
+ file_name,
+ input_body.get("content_type", DEFAULT_CONTENT_TYPE),
+ input_body.get("expiration", 60 * 60),
+ )
output = {
"message": "The S3 presigned url is generated",
"data": {
@@ -163,7 +179,6 @@ def lambda_handler(event, context):
"s3Bucket": s3_bucket_name,
"s3Prefix": file_name,
},
-
}
elif resource.startswith(EXECUTION_RESOURCE):
if http_method == "POST":
@@ -205,7 +220,7 @@ def __delete_execution(event, group_name):
"intentionId": execution_id,
},
)
- item = index_response.get('Item')
+ item = index_response.get("Item")
if item:
indexes = item.get("index").split(",")
details = json.loads(item.get("details"))
@@ -222,6 +237,7 @@ def __delete_execution(event, group_name):
)
return res
+
# def __can_be_deleted(execution_id):
# return False, ""
@@ -232,22 +248,21 @@ def __delete_documents_by_text_set(index_name, text_values):
search_body = {
"size": 10000,
"query": {
- "terms": {
- "text.keyword": list(text_values) # Convert set to list
- }
- }
+ "terms": {"text.keyword": list(text_values)} # Convert set to list
+ },
}
# Perform the search
try:
search_result = aos_client.search(
- index=index_name, body=search_body) # Adjust size if needed
- hits = search_result['hits']['hits']
+ index=index_name, body=search_body
+ ) # Adjust size if needed
+ hits = search_result["hits"]["hits"]
# If documents exist, delete them
if hits:
for hit in hits:
- doc_id = hit['_id']
+ doc_id = hit["_id"]
aos_client.delete(index=index_name, id=doc_id)
logger.info("Deleted document with id %s", doc_id)
except NotFoundError:
@@ -266,8 +281,11 @@ def __get_query_parameter(event, parameter_name, default_value=None):
def __gen_presigned_url(object_name: str, content_type: str, expiration: int):
return s3_client.generate_presigned_url(
ClientMethod="put_object",
- Params={"Bucket": s3_bucket_name,
- "Key": object_name, "ContentType": content_type},
+ Params={
+ "Bucket": s3_bucket_name,
+ "Key": object_name,
+ "ContentType": content_type,
+ },
ExpiresIn=expiration,
HttpMethod="PUT",
)
@@ -284,24 +302,20 @@ def __list_execution(event, group_name):
}
response = dynamodb_client.query(
TableName=intention_table_name,
- KeyConditionExpression='groupName = :groupName',
- ExpressionAttributeValues={
- ':groupName': {'S': group_name}
- }
+ KeyConditionExpression="groupName = :groupName",
+ ExpressionAttributeValues={":groupName": {"S": group_name}},
)
output = {}
page_json = []
- items = response['Items']
- while 'LastEvaluatedKey' in response:
+ items = response["Items"]
+ while "LastEvaluatedKey" in response:
response = dynamodb_client.query(
TableName=intention_table_name,
- KeyConditionExpression='groupName = :pk_val',
- ExpressionAttributeValues={
- ':pk_val': {'S': group_name}
- },
- ExclusiveStartKey=response['LastEvaluatedKey']
+ KeyConditionExpression="groupName = :pk_val",
+ ExpressionAttributeValues={":pk_val": {"S": group_name}},
+ ExclusiveStartKey=response["LastEvaluatedKey"],
)
- items.extend(response['Items'])
+ items.extend(response["Items"])
for item in items:
item_json = {}
@@ -321,7 +335,7 @@ def __list_execution(event, group_name):
item_json["executionStatus"] = value
else:
item_json[key] = value
-
+
page_json.append(item_json)
output["items"] = page_json
output["config"] = config
@@ -341,7 +355,7 @@ def __create_execution(event, context, email, group_name):
bucket = input_body.get("s3Bucket")
prefix = input_body.get("s3Prefix")
s3_response = __get_s3_object_with_retry(bucket, prefix)
- file_content = s3_response['Body'].read()
+ file_content = s3_response["Body"].read()
excel_file = BytesIO(file_content)
workbook = load_workbook(excel_file)
sheet = workbook.active
@@ -358,16 +372,21 @@ def __create_execution(event, context, email, group_name):
qd_index = chatbot_item.get("indexIds", {}).get("qd", {}).get("value", {})
valid_qd_types = [*built_in_tools, *list(qd_index.keys())]
for row in sheet.iter_rows(min_row=2, values_only=True):
- question, intention, kwargs = row[0], row[1], row[2] if len(
- row) > 2 else None
+ question, intention, kwargs = (
+ row[0],
+ row[1],
+ row[2] if len(row) > 2 else None,
+ )
if not question:
continue
- qaList.append({
- "question": question,
- "intention": intention,
- "kwargs": kwargs,
- "is_valid": intention in valid_qd_types
- })
+ qaList.append(
+ {
+ "question": question,
+ "intention": intention,
+ "kwargs": kwargs,
+ "is_valid": intention in valid_qd_types,
+ }
+ )
valid_qa_list = [qa for qa in qaList if qa.get("is_valid")]
# write to ddb(meta data)
@@ -381,20 +400,27 @@ def __create_execution(event, context, email, group_name):
"tag": execution_detail["index"],
"file": f'{bucket}{input_body.get("s3Prefix")}',
"lastModifiedBy": email,
- "lastModifiedTime": re.findall(r'\[(.*?)\]', input_body.get("s3Prefix"))[0],
+ "lastModifiedTime": re.findall(
+ r"\[(.*?)\]", input_body.get("s3Prefix")
+ )[0],
"details": json.dumps(qaList),
"validRatio": f"{len(valid_qa_list)} / {len(qaList)}",
}
)
-
+
# write to aos(vectorData)
- __save_2_aos(input_body.get("model"),
- execution_detail["index"], valid_qa_list, bucket, prefix)
+ __save_2_aos(
+ input_body.get("model"),
+ execution_detail["index"],
+ valid_qa_list,
+ bucket,
+ prefix,
+ )
return {
"execution_id": execution_detail["tableItemId"],
"input_payload": execution_detail,
- "result": "success"
+ "result": "success",
}
@@ -409,7 +435,7 @@ def convert_qa_list(qa_list: list, bucket: str, prefix: str) -> List[Document]:
"file_path": "",
"keywords": [],
"summary": "",
- "type": "Intent"
+ "type": "Intent",
}
page_content = qa["question"]
metadata = metadata_template
@@ -426,9 +452,11 @@ def convert_qa_list(qa_list: list, bucket: str, prefix: str) -> List[Document]:
return doc_list
-def __save_2_aos(modelId: str, index: str, qaListParam: list, bucket: str, prefix: str):
+def __save_2_aos(
+ modelId: str, index: str, qaListParam: list, bucket: str, prefix: str
+):
qaList = __deduplicate_by_key(qaListParam, "question")
- if kb_enabled and embedding_model_endpoint.startswith("bce-embedding"):
+ if kb_enabled:
embedding_info = get_embedding_info(embedding_model_endpoint)
embedding_function = sm_utils.getCustomEmbeddings(
embedding_model_endpoint, region, embedding_info.get("ModelType")
@@ -460,18 +488,14 @@ def __create_index(index: str, modelId: str):
"number_of_shards": 1,
"number_of_replicas": 0,
"knn": True,
- "knn.algo_param.ef_search": 32
+ "knn.algo_param.ef_search": 32,
}
},
"mappings": {
"properties": {
"text": {
"type": "text",
- "fields": {
- "keyword": {
- "type": "keyword"
- }
- }
+ "fields": {"keyword": {"type": "keyword"}},
},
"sentence_vector": {
"type": "knn_vector",
@@ -480,14 +504,11 @@ def __create_index(index: str, modelId: str):
"engine": "nmslib",
"space_type": "l2",
"name": "hnsw",
- "parameters": {
- "ef_construction": 512,
- "m": 16
- }
- }
- }
+ "parameters": {"ef_construction": 512, "m": 16},
+ },
+ },
}
- }
+ },
}
try:
aos_client.indices.create(index=index, body=body)
@@ -497,8 +518,11 @@ def __create_index(index: str, modelId: str):
def __refresh_index(index: str, modelId: str, qaList):
- success, failed = helpers.bulk(aos_client, __append_embeddings(
- index, modelId, qaList), chunk_size=BULK_SIZE)
+ success, failed = helpers.bulk(
+ aos_client,
+ __append_embeddings(index, modelId, qaList),
+ chunk_size=BULK_SIZE,
+ )
aos_client.indices.refresh(index=index)
logger.info("Successfully added: %d ", success)
logger.info("Failed: %d ", len(failed))
@@ -510,40 +534,40 @@ def __append_embeddings(index, modelId, qaList: list):
for item in qaList:
question = item["question"]
embedding_func = BedrockEmbeddings(
- client=bedrock_client,
- model_id=modelId,
- normalize=True
+ client=bedrock_client, model_id=modelId, normalize=True
)
- embeddings_vectors = embedding_func.embed_documents(
- [question]
- )
+ embeddings_vectors = embedding_func.embed_documents([question])
documents.append(
{
"text": question,
"metadata": {
"answer": item["intention"],
"source": "portal",
- **({"kwargs": item["kwargs"]} if item.get("kwargs") else {}),
- "type": "Intent"
+ **(
+ {"kwargs": item["kwargs"]} if item.get("kwargs") else {}
+ ),
+ "type": "Intent",
},
- "sentence_vector": embeddings_vectors[0]
+ "sentence_vector": embeddings_vectors[0],
}
)
for document in documents:
index_list = index.split(",")
for index_item in index_list:
- doc_id = hashlib.md5(str(document["text"]).encode('utf-8')).hexdigest()
+ doc_id = hashlib.md5(
+ str(document["text"]).encode("utf-8")
+ ).hexdigest()
action = {
"_op_type": "index",
"_index": index_item,
"_id": doc_id,
- "_source": document
+ "_source": document,
}
actions.append(action)
return actions
- # yield {"_op_type": "index", "_index": index_item, "_source": document, "_id": hashlib.md5(str(document).encode('utf-8')).hexdigest()}
+ # yield {"_op_type": "index", "_index": index_item, "_source": document, "_id": hashlib.md5(str(document).encode('utf-8')).hexdigest()}
def __get_execution(event, group_name):
@@ -554,7 +578,7 @@ def __get_execution(event, group_name):
"intentionId": executionId,
},
)
- item = index_response['Item']
+ item = index_response["Item"]
res = {}
items = []
# for item in items:
@@ -562,10 +586,10 @@ def __get_execution(event, group_name):
for key in list(item.keys()):
value = item.get(key)
if key == "file" or key == "File":
- split_index = value.rfind('/')
+ split_index = value.rfind("/")
if split_index != -1:
item_json["s3Path"] = value[:split_index]
- item_json["s3Prefix"] = value[split_index + 1:]
+ item_json["s3Prefix"] = value[split_index + 1 :]
else:
item_json["s3Path"] = value
item_json["s3Prefix"] = "-"
@@ -582,7 +606,9 @@ def __get_execution(event, group_name):
return res
-def __get_s3_object_with_retry(bucket: str, key: str, max_retries: int = 5, delay: int = 1):
+def __get_s3_object_with_retry(
+ bucket: str, key: str, max_retries: int = 5, delay: int = 1
+):
attempt = 0
while attempt < max_retries:
try:
@@ -595,12 +621,15 @@ def __get_s3_object_with_retry(bucket: str, key: str, max_retries: int = 5, dela
raise
time.sleep(delay)
+
def __download_template():
url = s3_client.generate_presigned_url(
ClientMethod="get_object",
- Params={'Bucket': s3_bucket_name,
- 'Key': "templates/intention_corpus.xlsx"},
- ExpiresIn=60
+ Params={
+ "Bucket": s3_bucket_name,
+ "Key": "templates/intention_corpus.xlsx",
+ },
+ ExpiresIn=60,
)
return url
@@ -614,7 +643,7 @@ def __index_used_scan(event, group_name):
},
)
pre_model = index_response.get("Item")
- model_name = ''
+ model_name = ""
if pre_model:
model_response = model_table.get_item(
Key={
@@ -622,25 +651,24 @@ def __index_used_scan(event, group_name):
"modelId": pre_model.get("modelIds", {}).get("embedding"),
}
)
- model_name = model_response.get("Item", {}).get(
- "parameter", {}).get("ModelName", "")
+ model_name = (
+ model_response.get("Item", {})
+ .get("parameter", {})
+ .get("ModelName", "")
+ )
# model_name = model_response.get("ModelName", {}).get("S","-")
if not pre_model or model_name == input_body.get("model"):
return {
"statusCode": 200,
"headers": resp_header,
- "body": json.dumps({
- "result": "valid"
- })
+ "body": json.dumps({"result": "valid"}),
}
else:
return {
"statusCode": 200,
"headers": resp_header,
- "body": json.dumps({
- "result": "invalid"
- }
- )}
+ "body": json.dumps({"result": "invalid"}),
+ }
def __deduplicate_by_key(lst, key):
@@ -649,6 +677,7 @@ def __deduplicate_by_key(lst, key):
seen[element[key]] = element
return list(seen.values())
+
def __get_query_parameter(event, parameter_name, default_value=None):
if (
event.get("queryStringParameters")
From 42552e9a509f04c35798b4b072446e39d9ddd504 Mon Sep 17 00:00:00 2001
From: Xu Han <nciiproto@gmail.com>
Date: Mon, 6 Jan 2025 02:09:25 +0000
Subject: [PATCH 2/6] fix: fix npm run build when aws configure not set
---
source/model/etl/code/model.sh | 6 +++---
source/script/build.sh | 3 ++-
2 files changed, 5 insertions(+), 4 deletions(-)
diff --git a/source/model/etl/code/model.sh b/source/model/etl/code/model.sh
index ea7b3886..0cff8011 100755
--- a/source/model/etl/code/model.sh
+++ b/source/model/etl/code/model.sh
@@ -6,16 +6,16 @@
dockerfile=$1
image=$2
tag=$3 # New argument for the tag
+aws_region=$4
-if [ "$image" = "" ] || [ "$dockerfile" = "" ] || [ "$tag" = "" ]
+if [ "$image" = "" ] || [ "$dockerfile" = "" ] || [ "$tag" = "" ] || [ "$aws_region" = "" ]
then
- echo "Usage: \$0 <docker-file> <image-name> <tag>"
+ echo "Usage: \$0 <docker-file> <image-name> <tag> <aws-region>"
exit 1
fi
# Get the account number associated with the current IAM credentials
account=$(aws sts get-caller-identity --query Account --output text)
-aws_region=$(aws configure get region)
echo "Account: $account"
echo "Region: $aws_region"
diff --git a/source/script/build.sh b/source/script/build.sh
index 70f0835f..d9931af9 100644
--- a/source/script/build.sh
+++ b/source/script/build.sh
@@ -4,6 +4,7 @@ set -e
# Load config.json
config_file="../infrastructure/bin/config.json"
+deploy_region=$(jq -r '.deployRegion' $config_file)
knowledge_base_enabled=$(jq -r '.knowledgeBase.enabled' $config_file)
knowledge_base_intelliagent_enabled=$(jq -r '.knowledgeBase.knowledgeBaseType.intelliAgentKb.enabled' $config_file)
knowledge_base_models_enabled=$(jq -r '.knowledgeBase.knowledgeBaseType.intelliAgentKb.knowledgeBaseModel.enabled' $config_file)
@@ -31,7 +32,7 @@ aws ecr-public get-login-password --region us-east-1 | docker login --username A
prepare_etl_model() {
echo "Preparing ETL Model"
cd model/etl/code
- sh model.sh ./Dockerfile $ecr_repository $ecr_image_tag
+ sh model.sh ./Dockerfile $ecr_repository $ecr_image_tag $deploy_region
cd - > /dev/null
pwd
}
From 7c74364fc59cff4762a8ce5eccf1815bb7cd9c01 Mon Sep 17 00:00:00 2001
From: zhouxss <zhouxss@amazon.com>
Date: Mon, 6 Jan 2025 05:28:40 +0000
Subject: [PATCH 3/6] fix dmaa integration bugs
---
.../lib/api/model-management.ts | 2 +
.../infrastructure/lib/shared/iam-helper.ts | 56 ++++++++++++++++++-
.../lambda/model_management/requirements.txt | 2 +-
.../chat_models/dmaa_models.py | 9 ++-
.../test/main_local_test_common.py | 24 ++++----
source/script/build.sh | 2 +-
6 files changed, 77 insertions(+), 18 deletions(-)
diff --git a/source/infrastructure/lib/api/model-management.ts b/source/infrastructure/lib/api/model-management.ts
index 39f5f0f4..feca62ec 100644
--- a/source/infrastructure/lib/api/model-management.ts
+++ b/source/infrastructure/lib/api/model-management.ts
@@ -66,6 +66,8 @@ export class ModelApi extends Construct {
modelLambda.addToRolePolicy(this.iamHelper.cfnStatement);
modelLambda.addToRolePolicy(this.iamHelper.stsStatement);
modelLambda.addToRolePolicy(this.iamHelper.cfnStatement);
+ modelLambda.addToRolePolicy(this.iamHelper.serviceQuotaStatement);
+ modelLambda.addToRolePolicy(this.iamHelper.sagemakerModelManagementStatement);
// API Gateway Lambda Integration to manage model
const lambdaModelIntegration = new apigw.LambdaIntegration(modelLambda, {
diff --git a/source/infrastructure/lib/shared/iam-helper.ts b/source/infrastructure/lib/shared/iam-helper.ts
index cf2c9562..c841bd48 100644
--- a/source/infrastructure/lib/shared/iam-helper.ts
+++ b/source/infrastructure/lib/shared/iam-helper.ts
@@ -30,6 +30,8 @@ export class IAMHelper extends Construct {
public secretStatement: PolicyStatement;
public codePipelineStatement: PolicyStatement;
public cfnStatement: PolicyStatement;
+ public serviceQuotaStatement: PolicyStatement;
+ public sagemakerModelManagementStatement: PolicyStatement;
public createPolicyStatement(actions: string[], resources: string[]) {
return new PolicyStatement({
@@ -79,7 +81,7 @@ export class IAMHelper extends Construct {
"sagemaker:CreateEndpoint",
"sagemaker:CreateEndpointConfig",
"sagemaker:InvokeEndpointAsync",
- "sagemaker:UpdateEndpointWeightsAndCapacities",
+ "sagemaker:UpdateEndpointWeightsAndCapacities"
],
[`arn:${Aws.PARTITION}:sagemaker:${Aws.REGION}:${Aws.ACCOUNT_ID}:endpoint/*`],
);
@@ -191,5 +193,57 @@ export class IAMHelper extends Construct {
],
["*"],
);
+ this.sagemakerModelManagementStatement = this.createPolicyStatement(
+ [
+ "sagemaker:List*",
+ "sagemaker:ListEndpoints",
+ "sagemaker:DeleteModel",
+ "sagemaker:DeleteEndpoint",
+ "sagemaker:DescribeEndpoint",
+ "sagemaker:DeleteEndpointConfig",
+ "sagemaker:DescribeEndpointConfig",
+ "sagemaker:InvokeEndpoint",
+ "sagemaker:CreateModel",
+ "sagemaker:CreateEndpoint",
+ "sagemaker:CreateEndpointConfig",
+ "sagemaker:InvokeEndpointAsync",
+ "sagemaker:UpdateEndpointWeightsAndCapacities"
+ ],
+ ["*"],
+ );
+ this.serviceQuotaStatement = this.createPolicyStatement(
+ [
+ "autoscaling:DescribeAccountLimits",
+ "cloudformation:DescribeAccountLimits",
+ "cloudwatch:DescribeAlarmsForMetric",
+ "cloudwatch:DescribeAlarms",
+ "cloudwatch:GetMetricData",
+ "cloudwatch:GetMetricStatistics",
+ "dynamodb:DescribeLimits",
+ "elasticloadbalancing:DescribeAccountLimits",
+ "iam:GetAccountSummary",
+ "kinesis:DescribeLimits",
+ "organizations:DescribeAccount",
+ "organizations:DescribeOrganization",
+ "organizations:ListAWSServiceAccessForOrganization",
+ "rds:DescribeAccountAttributes",
+ "route53:GetAccountLimit",
+ "tag:GetTagKeys",
+ "tag:GetTagValues",
+ "servicequotas:GetAssociationForServiceQuotaTemplate",
+ "servicequotas:GetAWSDefaultServiceQuota",
+ "servicequotas:GetRequestedServiceQuotaChange",
+ "servicequotas:GetServiceQuota",
+ "servicequotas:GetServiceQuotaIncreaseRequestFromTemplate",
+ "servicequotas:ListAWSDefaultServiceQuotas",
+ "servicequotas:ListRequestedServiceQuotaChangeHistory",
+ "servicequotas:ListRequestedServiceQuotaChangeHistoryByQuota",
+ "servicequotas:ListServices",
+ "servicequotas:ListServiceQuotas",
+ "servicequotas:ListServiceQuotaIncreaseRequestsInTemplate",
+ "servicequotas:ListTagsForResource"
+ ],
+ ["*"],
+ );
}
}
diff --git a/source/lambda/model_management/requirements.txt b/source/lambda/model_management/requirements.txt
index ab55f6ef..b5bbb176 100644
--- a/source/lambda/model_management/requirements.txt
+++ b/source/lambda/model_management/requirements.txt
@@ -1 +1 @@
-https://aws-gcr-solutions-assets.s3.us-east-1.amazonaws.com/dmaa/wheels/dmaa-0.4.0-py3-none-any.whl
+https://aws-gcr-solutions-assets.s3.us-east-1.amazonaws.com/dmaa/wheels/dmaa-0.5.0-py3-none-any.whl
diff --git a/source/lambda/online/common_logic/langchain_integration/chat_models/dmaa_models.py b/source/lambda/online/common_logic/langchain_integration/chat_models/dmaa_models.py
index 5f5d1958..70f82ea1 100644
--- a/source/lambda/online/common_logic/langchain_integration/chat_models/dmaa_models.py
+++ b/source/lambda/online/common_logic/langchain_integration/chat_models/dmaa_models.py
@@ -5,17 +5,20 @@
ModelProvider
)
import os
+import boto3
from dmaa.integrations.langchain_clients import SageMakerVllmChatModel as _SageMakerVllmChatModel
+session = boto3.Session()
+current_region = session.region_name
class SageMakerVllmChatModel(_SageMakerVllmChatModel):
- enable_auto_tool_choice: bool = True
+ enable_auto_tool_choice: bool = False
enable_prefill: bool = True
class Qwen25Instruct72bAwq(Model):
model_id = LLMModelType.QWEN25_INSTRUCT_72B_AWQ
- enable_auto_tool_choice: bool = True
+ enable_auto_tool_choice: bool = False
enable_prefill: bool = True
default_model_kwargs = {
"max_tokens": 2000,
@@ -32,7 +35,7 @@ def create_model(cls, model_kwargs=None, **kwargs):
or os.environ.get("AWS_PROFILE", None)
or None
)
- region_name = kwargs.get("region_name", None)
+ region_name = kwargs.get("region_name", None) or current_region
group_name = kwargs.get("group_name", os.environ.get('GROUP_NAME',"Admin"))
llm = SageMakerVllmChatModel(
diff --git a/source/lambda/online/lambda_main/test/main_local_test_common.py b/source/lambda/online/lambda_main/test/main_local_test_common.py
index 052ef7e4..ca73c178 100644
--- a/source/lambda/online/lambda_main/test/main_local_test_common.py
+++ b/source/lambda/online/lambda_main/test/main_local_test_common.py
@@ -167,20 +167,20 @@ def test_multi_turns_agent_pr():
# "query": "What does 245346356356 times 346357457 equal?", # 1089836033535
# "query": "9.11和9.9哪个更大？", # 1089836033535
# "query": "what happened in the history of Morgan Stanley in Emerging market in 1989 ?",
- "query": "Tell me the team members of Morgan Stanley in China",
- # "query": "今天天气如何？",
+ # "query": "Tell me the team members of Morgan Stanley in China",
+ "query": "今天天气如何？",
+ # "query": "介绍一下MemGPT",
+ "use_history": True,
+ "enable_trace": True
+ },
+ {
+ # "query": "”我爱北京天安门“包含多少个字符?",
+ # "query": "11133乘以97892395等于多少", # 1089836033535
+ "query": "我在上海",
# "query": "介绍一下MemGPT",
"use_history": True,
"enable_trace": True
},
- # {
- # # "query": "”我爱北京天安门“包含多少个字符?",
- # # "query": "11133乘以97892395等于多少", # 1089836033535
- # "query": "我在上海",
- # # "query": "介绍一下MemGPT",
- # "use_history": True,
- # "enable_trace": True
- # },
]
# default_index_names = {
@@ -193,9 +193,9 @@ def test_multi_turns_agent_pr():
# 'model_id': "anthropic.claude-3-5-sonnet-20240620-v1:0",
# 'model_id': "anthropic.claude-3-5-haiku-20241022-v1:0",
# 'model_id': "us.meta.llama3-2-90b-instruct-v1:0",
- 'model_id': "anthropic.claude-3-5-sonnet-20241022-v2:0",
+ # 'model_id': "anthropic.claude-3-5-sonnet-20241022-v2:0",
# 'model_id': "meta.llama3-1-70b-instruct-v1:0",
- # 'model_id':"mistral.mistral-large-2407-v1:0",
+ 'model_id':"mistral.mistral-large-2407-v1:0",
# 'model_id':"cohere.command-r-plus-v1:0",
'model_kwargs': {
'temperature': 0.01,
diff --git a/source/script/build.sh b/source/script/build.sh
index 70f0835f..6c126596 100644
--- a/source/script/build.sh
+++ b/source/script/build.sh
@@ -52,7 +52,7 @@ build_frontend() {
build_deployment_module() {
echo "Building Model Deployment Module"
- curl https://aws-gcr-solutions-assets.s3.us-east-1.amazonaws.com/dmaa/wheels/dmaa-0.4.0-py3-none-any.whl -o dmaa-0.4.0-py3-none-any.whl && pip install dmaa-0.4.0-py3-none-any.whl"[all]"
+ curl https://aws-gcr-solutions-assets.s3.us-east-1.amazonaws.com/dmaa/wheels/dmaa-0.5.0-py3-none-any.whl -o dmaa-0.5.0-py3-none-any.whl && pip install dmaa-0.5.0-py3-none-any.whl"[all]"
dmaa bootstrap
}
From 1d6d485fe063994696413c74c781f8a281119ad3 Mon Sep 17 00:00:00 2001
From: Xu Han <nciiproto@gmail.com>
Date: Tue, 7 Jan 2025 04:08:42 +0000
Subject: [PATCH 4/6] fix: fix etl change chatbot issue
---
source/lambda/etl/sfn_handler.py | 234 +++++++++++++++++++------------
1 file changed, 141 insertions(+), 93 deletions(-)
diff --git a/source/lambda/etl/sfn_handler.py b/source/lambda/etl/sfn_handler.py
index 94ec9b2e..c3bccf3b 100644
--- a/source/lambda/etl/sfn_handler.py
+++ b/source/lambda/etl/sfn_handler.py
@@ -2,124 +2,172 @@
import logging
import os
from datetime import datetime, timezone
+from typing import Dict, List, TypedDict
import boto3
-from chatbot_management import create_chatbot
from constant import ExecutionStatus, IndexType, UiStatus
from utils.parameter_utils import get_query_parameter
+# Initialize AWS resources once
client = boto3.client("stepfunctions")
dynamodb = boto3.resource("dynamodb")
execution_table = dynamodb.Table(os.environ.get("EXECUTION_TABLE_NAME"))
-index_table = dynamodb.Table(os.environ.get("INDEX_TABLE_NAME"))
chatbot_table = dynamodb.Table(os.environ.get("CHATBOT_TABLE_NAME"))
model_table = dynamodb.Table(os.environ.get("MODEL_TABLE_NAME"))
-embedding_endpoint = os.environ.get("EMBEDDING_ENDPOINT")
+index_table = dynamodb.Table(os.environ.get("INDEX_TABLE_NAME"))
sfn_arn = os.environ.get("SFN_ARN")
-create_time = str(datetime.now(timezone.utc))
+# Consolidate constants at the top
+CORS_HEADERS = {
+ "Content-Type": "application/json",
+ "Access-Control-Allow-Headers": "Content-Type,X-Amz-Date,Authorization,X-Api-Key,X-Amz-Security-Token",
+ "Access-Control-Allow-Origin": "*",
+ "Access-Control-Allow-Methods": "*",
+}
+
+# Initialize logging at the top level
logger = logging.getLogger()
logger.setLevel(logging.INFO)
-def handler(event, context):
- # Check the event for possible S3 created event
- input_payload = {}
- logger.info(event)
- resp_header = {
- "Content-Type": "application/json",
- "Access-Control-Allow-Headers": "Content-Type,X-Amz-Date,Authorization,X-Api-Key,X-Amz-Security-Token",
- "Access-Control-Allow-Origin": "*",
- "Access-Control-Allow-Methods": "*",
+def validate_index_type(index_type: str) -> bool:
+ """Validate if the provided index type is supported."""
+ valid_types = [
+ IndexType.QD.value,
+ IndexType.QQ.value,
+ IndexType.INTENTION.value,
+ ]
+ return index_type in valid_types
+
+
+def get_etl_info(group_name: str, chatbot_id: str, index_type: str):
+ """
+ Retrieve the index id, model type, and model endpoint for the given chatbot and index type.
+ These will be further used to perform knowledge ingestion to opensearch.
+ Returns: Tuple of (index_id, model_type, model_endpoint)
+ """
+
+ chatbot_item = chatbot_table.get_item(
+ Key={"groupName": group_name, "chatbotId": chatbot_id}
+ ).get("Item")
+
+ model_item = model_table.get_item(
+ Key={"groupName": group_name, "modelId": f"{chatbot_id}-embedding"}
+ ).get("Item")
+
+ if not (chatbot_item and model_item):
+ raise ValueError("Chatbot or model not found")
+
+ model = model_item.get("parameter", {})
+ specific_type_indices = (
+ chatbot_item.get("indexIds", {}).get(index_type, {}).get("value", {})
+ )
+
+ if not specific_type_indices:
+ raise ValueError("No indices found for the given index type")
+
+ return (
+ next(iter(specific_type_indices.values())), # First index ID
+ model.get("ModelType"),
+ model.get("ModelEndpoint"),
+ )
+
+
+def create_execution_record(
+ execution_id: str, input_body: Dict, sfn_execution_id: str
+) -> None:
+ """Create execution record in DynamoDB."""
+ execution_record = {
+ **input_body,
+ "sfnExecutionId": sfn_execution_id,
+ "executionStatus": ExecutionStatus.IN_PROGRESS.value,
+ "executionId": execution_id,
+ "uiStatus": UiStatus.ACTIVE.value,
+ "createTime": str(datetime.now(timezone.utc)),
}
+ del execution_record["tableItemId"]
+ execution_table.put_item(Item=execution_record)
+
+
+def handler(event: Dict, context) -> Dict:
+ """Main Lambda handler for ETL operations."""
+ logger.info(event)
- authorizer_type = event["requestContext"].get("authorizer", {}).get("authorizerType")
- if authorizer_type == "lambda_authorizer":
- claims = json.loads(event["requestContext"]["authorizer"]["claims"])
+ try:
+ # Validate and extract authorization
+ authorizer = event["requestContext"].get("authorizer", {})
+ if authorizer.get("authorizerType") != "lambda_authorizer":
+ raise ValueError("Invalid authorizer type")
+
+ claims = json.loads(authorizer.get("claims", {}))
if "use_api_key" in claims:
group_name = get_query_parameter(event, "GroupName", "Admin")
cognito_groups_list = [group_name]
else:
- cognito_groups = claims["cognito:groups"]
- cognito_groups_list = cognito_groups.split(",")
- else:
- logger.error("Invalid authorizer type")
+ cognito_groups_list = claims["cognito:groups"].split(",")
+
+ # Process input
+ input_body = json.loads(event["body"])
+ index_type = input_body.get("indexType")
+
+ if not validate_index_type(index_type):
+ return {
+ "statusCode": 400,
+ "headers": CORS_HEADERS,
+ "body": f"Invalid indexType, valid values are {', '.join([t.value for t in IndexType])}",
+ }
+
+ group_name = input_body.get("groupName") or (
+ "Admin"
+ if "Admin" in cognito_groups_list
+ else cognito_groups_list[0]
+ )
+ chatbot_id = input_body.get("chatbotId", group_name.lower())
+ index_id, embedding_model_type, embedding_endpoint = get_etl_info(
+ group_name, chatbot_id, index_type
+ )
+
+ # Update input body with processed values
+ input_body.update(
+ {
+ "chatbotId": chatbot_id,
+ "groupName": group_name,
+ "tableItemId": context.aws_request_id,
+ "indexId": index_id,
+ "embeddingModelType": embedding_model_type,
+ "embeddingEndpoint": embedding_endpoint,
+ }
+ )
+
+ # Start step function and create execution record
+ sfn_response = client.start_execution(
+ stateMachineArn=sfn_arn, input=json.dumps(input_body)
+ )
+
+ execution_id = context.aws_request_id
+ create_execution_record(
+ execution_id,
+ input_body,
+ sfn_response["executionArn"].split(":")[-1],
+ )
+
return {
- "statusCode": 403,
- "headers": resp_header,
- "body": json.dumps({"error": "Invalid authorizer type"}),
+ "statusCode": 200,
+ "headers": CORS_HEADERS,
+ "body": json.dumps(
+ {
+ "execution_id": execution_id,
+ "step_function_arn": sfn_response["executionArn"],
+ "input_payload": input_body,
+ }
+ ),
}
- # Parse the body from the event object
- input_body = json.loads(event["body"])
- if "indexType" not in input_body or input_body["indexType"] not in [
- IndexType.QD.value,
- IndexType.QQ.value,
- IndexType.INTENTION.value,
- ]:
+ except Exception as e:
+ logger.error(f"Error processing request: {str(e)}")
return {
- "statusCode": 400,
- "headers": resp_header,
- "body": (
- f"Invalid indexType, valid values are "
- f"{IndexType.QD.value}, {IndexType.QQ.value}, "
- f"{IndexType.INTENTION.value}"
- ),
+ "statusCode": 500,
+ "headers": CORS_HEADERS,
+ "body": json.dumps({"error": str(e)}),
}
- index_type = input_body["indexType"]
- group_name = "Admin" if "Admin" in cognito_groups_list else cognito_groups_list[0]
- chatbot_id = input_body.get("chatbotId", group_name.lower())
-
- if "indexId" in input_body:
- index_id = input_body["indexId"]
- else:
- # Use default index id if not specified in the request
- index_id = f"{chatbot_id}-qd-default"
- if index_type == IndexType.QQ.value:
- index_id = f"{chatbot_id}-qq-default"
- elif index_type == IndexType.INTENTION.value:
- index_id = f"{chatbot_id}-intention-default"
-
- if "tag" in input_body:
- tag = input_body["tag"]
- else:
- tag = index_id
-
- input_body["indexId"] = index_id
- input_body["groupName"] = group_name if "groupName" not in input_body else input_body["groupName"]
- chatbot_event_body = input_body
- chatbot_event_body["group_name"] = group_name
- chatbot_event = {"body": json.dumps(chatbot_event_body)}
- chatbot_result = create_chatbot(chatbot_event, group_name)
-
- input_body["tableItemId"] = context.aws_request_id
- input_body["chatbotId"] = chatbot_id
- input_body["embeddingModelType"] = chatbot_result["modelType"]
- input_payload = json.dumps(input_body)
- response = client.start_execution(stateMachineArn=sfn_arn, input=input_payload)
-
- # Update execution table item
- if "tableItemId" in input_body:
- del input_body["tableItemId"]
- execution_id = response["executionArn"].split(":")[-1]
- input_body["sfnExecutionId"] = execution_id
- input_body["executionStatus"] = ExecutionStatus.IN_PROGRESS.value
- input_body["indexId"] = index_id
- input_body["executionId"] = context.aws_request_id
- input_body["uiStatus"] = UiStatus.ACTIVE.value
- input_body["createTime"] = create_time
-
- execution_table.put_item(Item=input_body)
-
- return {
- "statusCode": 200,
- "headers": resp_header,
- "body": json.dumps(
- {
- "execution_id": context.aws_request_id,
- "step_function_arn": response["executionArn"],
- "input_payload": input_payload,
- }
- ),
- }
From 5f706af1d21059d703ce259ea8737e5a284c918f Mon Sep 17 00:00:00 2001
From: Cui <530051970@qq.com>
Date: Tue, 7 Jan 2025 16:21:36 +0800
Subject: [PATCH 5/6] bug fix: input component issues when pages opened by
firefox
---
source/portal/package-lock.json | 27 ++++++++---------
source/portal/package.json | 1 -
source/portal/src/pages/chatbot/ChatBot.tsx | 33 ++++++++++++++-------
3 files changed, 35 insertions(+), 26 deletions(-)
diff --git a/source/portal/package-lock.json b/source/portal/package-lock.json
index ef678152..2eb5c8b3 100644
--- a/source/portal/package-lock.json
+++ b/source/portal/package-lock.json
@@ -8,7 +8,6 @@
"name": "portal",
"version": "0.0.0",
"dependencies": {
- "@cloudscape-design/component-toolkit": "^1.0.0-beta.81",
"@cloudscape-design/components": "^3.0.604",
"@cloudscape-design/global-styles": "^1.0.27",
"axios": "^1.7.4",
@@ -444,18 +443,18 @@
}
},
"node_modules/@cloudscape-design/component-toolkit": {
- "version": "1.0.0-beta.81",
- "resolved": "https://registry.npmjs.org/@cloudscape-design/component-toolkit/-/component-toolkit-1.0.0-beta.81.tgz",
- "integrity": "sha512-//WS5C+DSi6vbD17gdyUO9hikFWjID8FXSmzqtAxV6FsO2HRiSJkAMLt6jRCMZKo9JpOnDaEhtx5pkuiY/Ez+g==",
+ "version": "1.0.0-beta.47",
+ "resolved": "https://registry.npmjs.org/@cloudscape-design/component-toolkit/-/component-toolkit-1.0.0-beta.47.tgz",
+ "integrity": "sha512-A2rJJo7/OBosTKLEZ75k92/yKB6nSWgbGZCfIFadm/Y+lYiW0lShU98B4MnVqGqfQigbNaBtlChKn4Aixxk+qg==",
"dependencies": {
"@juggle/resize-observer": "^3.3.1",
"tslib": "^2.3.1"
}
},
"node_modules/@cloudscape-design/components": {
- "version": "3.0.857",
- "resolved": "https://registry.npmjs.org/@cloudscape-design/components/-/components-3.0.857.tgz",
- "integrity": "sha512-v0vwY/PsPHQF62mE5e0nqV7xJdkHbITgFUNIKoVcJtx10SzomKRBwn4+odWZKo6E6LU8paeB2vTPdXgrOJtpnw==",
+ "version": "3.0.604",
+ "resolved": "https://registry.npmjs.org/@cloudscape-design/components/-/components-3.0.604.tgz",
+ "integrity": "sha512-PtYv3sp0SbgGmh7LgTjTmnN26yGkFwa7yMSea0qXrbLhIJ//eXC0Mj5Kp6pzBaZ1U+4e40km4bWv/EASVQK8Ag==",
"dependencies": {
"@cloudscape-design/collection-hooks": "^1.0.0",
"@cloudscape-design/component-toolkit": "^1.0.0-beta",
@@ -465,7 +464,7 @@
"@dnd-kit/sortable": "^7.0.2",
"@dnd-kit/utilities": "^3.2.1",
"@juggle/resize-observer": "^3.3.1",
- "ace-builds": "1.36.0",
+ "ace-builds": "^1.32.6",
"balanced-match": "^1.0.2",
"clsx": "^1.1.0",
"d3-shape": "^1.3.7",
@@ -498,9 +497,9 @@
}
},
"node_modules/@cloudscape-design/global-styles": {
- "version": "1.0.33",
- "resolved": "https://registry.npmjs.org/@cloudscape-design/global-styles/-/global-styles-1.0.33.tgz",
- "integrity": "sha512-6bg18XIxkRS2ojMNGxVA8mV35rqkiHDXwOJjfHhYPzg6LjFagZWyg/hRRGuP5MExszB748m2HYYdXT0EejxiPA=="
+ "version": "1.0.27",
+ "resolved": "https://registry.npmjs.org/@cloudscape-design/global-styles/-/global-styles-1.0.27.tgz",
+ "integrity": "sha512-26/Xt6eVB+xV+WCZJVmckiUUCYKrhQdGIr1t5gzpO/2VvERqjf+8x3buznyxNlFEfXaiKTMBtATnnYe27ARSiw=="
},
"node_modules/@cloudscape-design/test-utils-core": {
"version": "1.0.30",
@@ -1792,9 +1791,9 @@
}
},
"node_modules/ace-builds": {
- "version": "1.36.0",
- "resolved": "https://registry.npmjs.org/ace-builds/-/ace-builds-1.36.0.tgz",
- "integrity": "sha512-7to4F86V5N13EY4M9LWaGo2Wmr9iWe5CrYpc28F+/OyYCf7yd+xBV5x9v/GB73EBGGoYd89m6JjeIUjkL6Yw+w=="
+ "version": "1.32.9",
+ "resolved": "https://registry.npmjs.org/ace-builds/-/ace-builds-1.32.9.tgz",
+ "integrity": "sha512-dqBLPj//Gq0b92YUtRIsdWsORf4J+4xW3r8/4Wr2Vqid7O1j7YBV/ZsVvWBjZFy+EnvMCRFCFOEIM1cbt4BQ/g=="
},
"node_modules/acorn": {
"version": "8.11.3",
diff --git a/source/portal/package.json b/source/portal/package.json
index 8d3a5302..c020eb62 100644
--- a/source/portal/package.json
+++ b/source/portal/package.json
@@ -10,7 +10,6 @@
"preview": "vite preview"
},
"dependencies": {
- "@cloudscape-design/component-toolkit": "^1.0.0-beta.81",
"@cloudscape-design/components": "^3.0.604",
"@cloudscape-design/global-styles": "^1.0.27",
"axios": "^1.7.4",
diff --git a/source/portal/src/pages/chatbot/ChatBot.tsx b/source/portal/src/pages/chatbot/ChatBot.tsx
index 6c98bf84..0bb58c37 100644
--- a/source/portal/src/pages/chatbot/ChatBot.tsx
+++ b/source/portal/src/pages/chatbot/ChatBot.tsx
@@ -392,13 +392,16 @@ const ChatBot: React.FC<ChatBotProps> = (props: ChatBotProps) => {
}
};
- document.addEventListener('compositionstart', () => {
- setIsComposing(true);
- });
+ const inputElement = document.querySelector('input');
- document.addEventListener('compositionend', () => {
- setIsComposing(false);
- });
+ if (inputElement) {
+ inputElement.addEventListener('compositionstart', () => {
+ setIsComposing(true);
+ });
+ inputElement.addEventListener('compositionend', () => {
+ setIsComposing(false);
+ });
+ }
useEffect(() => {
if (lastMessage !== null) {
@@ -467,23 +470,25 @@ const ChatBot: React.FC<ChatBotProps> = (props: ChatBotProps) => {
}
if (!maxRounds.trim()) {
- setMaxTokenError('validation.requireMaxRounds');
+ setMaxRoundsError('validation.requireMaxRounds');
setModelSettingExpand(true);
return;
}
+
if (parseInt(maxRounds) < 0) {
- setMaxTokenError('validation.maxRoundsRange');
+ setMaxRoundsError('validation.maxRoundsRange');
setModelSettingExpand(true);
return;
}
if (!topKRetrievals.trim()) {
- setMaxTokenError('validation.requireTopKRetrievals');
+ setTopKRetrievalsError('validation.requireTopKRetrievals');
setModelSettingExpand(true);
return;
}
+
if (parseInt(topKRetrievals) < 1) {
- setMaxTokenError('validation.topKRetrievals');
+ setTopKRetrievalsError('validation.topKRetrievals');
setModelSettingExpand(true);
return;
}
@@ -790,6 +795,9 @@ const ChatBot: React.FC<ChatBotProps> = (props: ChatBotProps) => {
type="number"
value={maxRounds}
onChange={({ detail }) => {
+ if(parseInt(detail.value) < 0 || parseInt(detail.value) > 100){
+ return
+ }
setMaxRoundsError('');
setMaxRounds(detail.value);
}}
@@ -851,6 +859,9 @@ const ChatBot: React.FC<ChatBotProps> = (props: ChatBotProps) => {
type="number"
value={topKRetrievals}
onChange={({ detail }) => {
+ if(parseInt(detail.value) < 0 || parseInt(detail.value) > 100){
+ return
+ }
setTopKRetrievalsError('');
setTopKRetrievals(detail.value);
}}
@@ -864,7 +875,7 @@ const ChatBot: React.FC<ChatBotProps> = (props: ChatBotProps) => {
>
<Input
type="number"
- step={0.1}
+ step={0.01}
value={score}
onChange={({ detail }) => {
if(parseFloat(detail.value) < 0 || parseFloat(detail.value) > 1){
From 70163d32d5d6da3ba80277e3f7e17247b7897f97 Mon Sep 17 00:00:00 2001
From: Cui <530051970@qq.com>
Date: Tue, 7 Jan 2025 16:31:49 +0800
Subject: [PATCH 6/6] clear user messages when start a new chat
---
source/portal/src/pages/chatbot/ChatBot.tsx | 1 +
1 file changed, 1 insertion(+)
diff --git a/source/portal/src/pages/chatbot/ChatBot.tsx b/source/portal/src/pages/chatbot/ChatBot.tsx
index 0bb58c37..fe2cb748 100644
--- a/source/portal/src/pages/chatbot/ChatBot.tsx
+++ b/source/portal/src/pages/chatbot/ChatBot.tsx
@@ -171,6 +171,7 @@ const ChatBot: React.FC<ChatBotProps> = (props: ChatBotProps) => {
setTemperature(defaultConfig.temperature)
setTopKRetrievals(defaultConfig.topKRetrievals)
setScore(defaultConfig.score)
+ setUserMessage('')
setAdditionalConfig('')
// setModelOption(optionList?.[0]?.value ?? '')
setSessionId(uuidv4());