Bump llama_index version (#341)

* Update poetry * Fix poetry * bump llama_index to v0.12.3 * fix deps * Fix dropdown value * fix deps * generate poetry.lock * add html2text * add html2text * Fix test * fix tests --------- Co-authored-by: Yue Fei <[email protected]>
aigc-apps · Jan 16, 2025 · cefc173 · cefc173
1 parent 97c7b10
commit cefc173
Show file tree

Hide file tree

Showing 38 changed files with 2,265 additions and 2,709 deletions.
diff --git a/Dockerfile b/Dockerfile
@@ -24,7 +24,7 @@ ENV VIRTUAL_ENV=/app/.venv \
     ENABLE_AIOHTTPCLIENT=false \
     ENABLE_HTTPX=false
 
-RUN apt-get update && apt-get install -y libgl1 libglib2.0-0 curl
+RUN apt-get update && apt-get install -y libgl1 libglib2.0-0 libgomp1 curl
 
 # setup paddleocr dependencies
 RUN mkdir -p /root/.paddleocr/whl/det/ch/ch_PP-OCRv4_det_infer \

diff --git a/Dockerfile_gpu b/Dockerfile_gpu
@@ -27,7 +27,7 @@ ENV VIRTUAL_ENV=/app/.venv \
     ENABLE_AIOHTTPCLIENT=false \
     ENABLE_HTTPX=false
 
-RUN apt-get update && apt-get install -y libgl1 libglib2.0-0 curl
+RUN apt-get update && apt-get install -y libgl1 libglib2.0-0 libgomp1 curl
 
 # setup paddleocr dependencies
 RUN mkdir -p /root/.paddleocr/whl/det/ch/ch_PP-OCRv4_det_infer \

diff --git a/poetry.lock b/poetry.lock
diff --git a/pyproject.toml b/pyproject.toml
@@ -11,32 +11,39 @@ readme = "README.md"
 
 [tool.poetry.dependencies]
 python = ">=3.11.0,<3.12"
-fastapi = "^0.110.1"
+fastapi = "0.115.2"
 uvicorn = "^0.29.0"
-llama-index-core = "0.10.62"
-llama-index-embeddings-openai = "^0.1.7"
-llama-index-embeddings-azure-openai = "^0.1.7"
-llama-index-embeddings-dashscope = "^0.1.3"
-llama-index-llms-openai = "^0.1.27"
-llama-index-llms-azure-openai = "^0.1.6"
-llama-index-llms-dashscope = "^0.1.2"
-llama-index-readers-database = "^0.1.3"
-llama-index-vector-stores-faiss = "^0.1.2"
-llama-index-vector-stores-analyticdb = "^0.1.1"
-llama-index-vector-stores-elasticsearch = "^0.2.0"
-llama-index-vector-stores-milvus = "^0.1.10"
-gradio = "3.41.0"
+llama-index-core = "^0.12.2"
+llama-index-embeddings-openai = "^0.3.0"
+llama-index-embeddings-azure-openai = "^0.3.0"
+llama-index-embeddings-dashscope = "^0.3.0"
+llama-index-llms-openai = "^0.3.0"
+llama-index-llms-azure-openai = "^0.3.0"
+llama-index-llms-dashscope = "^0.3.0"
+llama-index-readers-database = "^0.3.0"
+llama-index-vector-stores-faiss = "^0.3.0"
+llama-index-vector-stores-analyticdb = "^0.3.0"
+llama-index-vector-stores-elasticsearch = "^0.4.0"
+llama-index-vector-stores-milvus = "^0.3.0"
+llama-index-vector-stores-alibabacloud-opensearch = "^0.3.0"
+llama-index-embeddings-huggingface = "^0.4.0"
+llama-index-readers-file = "^0.4.3"
+llama-index-postprocessor-flag-embedding-reranker = "^0.3.0"
+llama-index-storage-chat-store-redis = "^0.4.0"
+llama-index-tools-google = "^0.3.0"
+llama-index-tools-duckduckgo = "^0.3.0"
+llama-index-llms-paieas = "^0.3.0"
+llama-index-agent-openai = "^0.4.1"
+llama-index-multi-modal-llms-openai = "^0.4.2"
+gradio = "^5.12.0"
 faiss-cpu = "^1.8.0"
 hologres-vector = "^0.0.9"
 dynaconf = "^3.2.5"
 docx2txt = "^0.8"
 click = "^8.1.7"
 pydantic = "^2.7.0"
 pytest = "^8.1.1"
-llama-index-retrievers-bm25 = "^0.1.3"
 jieba = "^0.42.1"
-llama-index-embeddings-huggingface = "^0.2.0"
-llama-index-postprocessor-flag-embedding-reranker = "^0.1.3"
 flagembedding = "^1.2.10"
 sentencepiece = "^0.2.0"
 oss2 = "^2.18.5"
@@ -50,10 +57,9 @@ torchvision = [
   {version = "0.18.0+cpu", source = "pytorch_cpu", markers = "sys_platform != 'darwin'"},
   {version = "0.17.2", markers = "sys_platform == 'darwin'"}
 ]
-transformers = "4.42.4"
+transformers = "^4.42.4"
 openpyxl = "^3.1.2"
 pdf2image = "^1.17.0"
-llama-index-storage-chat-store-redis = "^0.1.3"
 python-bidi = "0.4.2"
 easyocr = "^1.7.1"
 opencv-python = "^4.6.0.66"
@@ -62,10 +68,7 @@ pypdf2 = "^3.0.1"
 pdfplumber = "^0.11.0"
 pdfminer-six = "^20231228"
 openinference-semantic-conventions = "^0.1.9"
-llama-index-tools-google = "^0.1.5"
-llama-index-tools-duckduckgo = "^0.1.1"
 openinference-instrumentation = "^0.1.12"
-llama-index-llms-huggingface = "^0.2.0"
 pytest-asyncio = "^0.23.7"
 pytest-cov = "^5.0.0"
 xlrd = "^2.0.1"
@@ -74,41 +77,37 @@ chardet = "^5.2.0"
 locust = "^2.29.0"
 gunicorn = "^22.0.0"
 umap-learn = "^0.5.6"
-protobuf = "3.20.2"
+protobuf = "^3.20.2"
 modelscope = "^1.16.0"
-llama-index-multi-modal-llms-dashscope = "^0.1.2"
-llama-index-vector-stores-alibabacloud-opensearch = "^0.1.0"
 asyncpg = "^0.29.0"
 pgvector = "^0.3.2"
 pre-commit = "^3.8.0"
 cn-clip = "^1.5.1"
-llama-index-llms-paieas = "^0.1.0"
 pymysql = "^1.1.1"
-llama-index-experimental = "^0.2.0"
-llama-index-readers-web = "^0.1.23"
 milvus-lite = "^2.4.9"
 rapidocr-onnxruntime = "^1.3.24"
 rapid-table = "^0.1.3"
 bs4 = "^0.0.2"
-httpx = "0.27.0"
+httpx = "^0.27.0"
 detectron2 = [
   {markers = "sys_platform == 'linux'", url = "https://pai-rag.oss-cn-hangzhou.aliyuncs.com/packages/python_wheels/detectron2-0.6%2B864913fpt2.3.0cpu-cp311-cp311-linux_x86_64.whl"},
   {markers = "sys_platform == 'win32'", url = "https://pai-rag.oss-cn-hangzhou.aliyuncs.com/packages/python_wheels/detectron2-0.6%2B864913fpt2.3.0cpu-cp311-cp311-win_amd64.whl"},
   {markers = "sys_platform != 'win32' and sys_platform != 'linux' ", url = "https://pai-rag.oss-cn-hangzhou.aliyuncs.com/packages/python_wheels/detectron2-0.6%2B864913fpt2.2.2cpu-cp311-cp311-macosx_10_9_universal2.whl"}
 ]
-magic-pdf = {extras = ["full"], version = "^0.10.6"}
+magic-pdf = {version = "0.10.6", extras = ["full"]}
 peft = "^0.12.0"
-duckduckgo-search = "6.2.12"
-aliyun-bootstrap = "1.0.2"
+duckduckgo-search = "^6.2.12"
+aliyun-bootstrap = "^1.0.2"
 docx = "^0.2.4"
 python-pptx = "^1.0.2"
 aspose-slides = "^24.10.0"
-ultralytics = "^8.3.58"
+ultralytics = "^8.3.43"
 datasketch = "^1.6.5"
-primp = "0.9.1"
+primp = "^0.9.1"
 tablestore = "^6.1.0"
-anyio = "4.6.2.post1"
+anyio = "^4.6.2.post1"
 mistletoe = "^1.4.0"
+html2text = "^2024.2.26"
 
 [tool.poetry.scripts]
 pai_rag = "pai_rag.main:run"

diff --git a/pyproject_gpu.toml b/pyproject_gpu.toml
@@ -11,32 +11,38 @@ readme = "README.md"
 
 [tool.poetry.dependencies]
 python = ">=3.11.0,<3.12"
-fastapi = "^0.110.1"
+fastapi = "0.115.2"
 uvicorn = "^0.29.0"
-llama-index-core = "0.10.62"
-llama-index-embeddings-openai = "^0.1.7"
-llama-index-embeddings-azure-openai = "^0.1.7"
-llama-index-embeddings-dashscope = "^0.1.3"
-llama-index-llms-openai = "^0.1.27"
-llama-index-llms-azure-openai = "^0.1.6"
-llama-index-llms-dashscope = "^0.1.2"
-llama-index-readers-database = "^0.1.3"
-llama-index-vector-stores-faiss = "^0.1.2"
-llama-index-vector-stores-analyticdb = "^0.1.1"
-llama-index-vector-stores-elasticsearch = "^0.2.0"
-llama-index-vector-stores-milvus = "^0.1.10"
-gradio = "3.41.0"
+llama-index-core = "^0.12.2"
+llama-index-embeddings-openai = "^0.3.0"
+llama-index-embeddings-azure-openai = "^0.3.0"
+llama-index-embeddings-dashscope = "^0.3.0"
+llama-index-llms-openai = "^0.3.0"
+llama-index-llms-azure-openai = "^0.3.0"
+llama-index-llms-dashscope = "^0.3.0"
+llama-index-readers-database = "^0.3.0"
+llama-index-vector-stores-faiss = "^0.3.0"
+llama-index-vector-stores-analyticdb = "^0.3.0"
+llama-index-vector-stores-elasticsearch = "^0.4.0"
+llama-index-vector-stores-milvus = "^0.3.0"
+llama-index-agent-openai = "^0.4.1"
+llama-index-embeddings-huggingface = "^0.4.0"
+llama-index-postprocessor-flag-embedding-reranker = "^0.3.0"
+llama-index-storage-chat-store-redis = "^0.4.0"
+llama-index-tools-google = "^0.3.0"
+llama-index-tools-duckduckgo = "^0.3.0"
+llama-index-multi-modal-llms-openai = "^0.4.2"
+llama-index-vector-stores-alibabacloud-opensearch = "^0.3.0"
+llama-index-llms-paieas = "^0.3.0"
+gradio = "^5.12.0"
 faiss-cpu = "^1.8.0"
 hologres-vector = "^0.0.9"
 dynaconf = "^3.2.5"
 docx2txt = "^0.8"
 click = "^8.1.7"
 pydantic = "^2.7.0"
 pytest = "^8.1.1"
-llama-index-retrievers-bm25 = "^0.1.3"
 jieba = "^0.42.1"
-llama-index-embeddings-huggingface = "^0.2.0"
-llama-index-postprocessor-flag-embedding-reranker = "^0.1.3"
 flagembedding = "^1.2.10"
 sentencepiece = "^0.2.0"
 oss2 = "^2.18.5"
@@ -47,7 +53,6 @@ torchvision = "0.17.2"
 transformers = "4.42.4"
 openpyxl = "^3.1.2"
 pdf2image = "^1.17.0"
-llama-index-storage-chat-store-redis = "^0.1.3"
 python-bidi = "0.4.2"
 easyocr = "^1.7.1"
 opencv-python = "^4.6.0.66"
@@ -56,10 +61,7 @@ pypdf2 = "^3.0.1"
 pdfplumber = "^0.11.0"
 pdfminer-six = "^20231228"
 openinference-semantic-conventions = "^0.1.9"
-llama-index-tools-google = "^0.1.5"
-llama-index-tools-duckduckgo = "^0.1.1"
 openinference-instrumentation = "^0.1.12"
-llama-index-llms-huggingface = "^0.2.0"
 pytest-asyncio = "^0.23.7"
 pytest-cov = "^5.0.0"
 xlrd = "^2.0.1"
@@ -68,18 +70,13 @@ chardet = "^5.2.0"
 locust = "^2.29.0"
 gunicorn = "^22.0.0"
 umap-learn = "^0.5.6"
-protobuf = "3.20.2"
+protobuf = "^3.20.2"
 modelscope = "^1.16.0"
-llama-index-multi-modal-llms-dashscope = "^0.1.2"
-llama-index-vector-stores-alibabacloud-opensearch = "^0.1.0"
 asyncpg = "^0.29.0"
 pgvector = "^0.3.2"
 pre-commit = "^3.8.0"
 cn-clip = "^1.5.1"
-llama-index-llms-paieas = "^0.1.0"
 pymysql = "^1.1.1"
-llama-index-experimental = "^0.2.0"
-llama-index-readers-web = "^0.1.23"
 milvus-lite = "^2.4.9"
 rapidocr-onnxruntime = "^1.3.24"
 rapid-table = "^0.1.3"
@@ -100,11 +97,12 @@ aliyun-bootstrap = "^1.0.2"
 docx = "^0.2.4"
 python-pptx = "^1.0.2"
 aspose-slides = "^24.10.0"
-ultralytics = "8.3.58"
+ultralytics = "^8.3.43"
 datasketch = "^1.6.5"
 tablestore = "^6.1.0"
+anyio = "^4.6.2.post1"
 mistletoe = "^1.4.0"
-anyio = "4.6.2.post1"
+html2text = "^2024.2.26"
 
 [tool.poetry.scripts]
 pai_rag = "pai_rag.main:run"

diff --git a/src/pai_rag/app/web/index_utils.py b/src/pai_rag/app/web/index_utils.py
@@ -180,7 +180,7 @@ def index_to_components_settings(
                 {"value": ""},
                 {"value": ""},
                 {"value": ""},
-                {"value": ""},
+                {"value": "cn-hangzhou"},
                 {"value": ""},
                 {"value": ""},
                 {"value": ""},

diff --git a/src/pai_rag/app/web/rag_client.py b/src/pai_rag/app/web/rag_client.py
@@ -208,6 +208,7 @@ def _format_rag_response(
             formatted_answer += f"**Reference**:\n {referenced_docs}"
 
         response["result"] = formatted_answer
+
         return response
 
     def check_health(self):

diff --git a/src/pai_rag/app/web/tabs/data_analysis_tab.py b/src/pai_rag/app/web/tabs/data_analysis_tab.py
@@ -57,7 +57,6 @@ def load_db_info_fn(input_elements: List[Any]):
         else:
             update_dict["enable_query_preprocessor"] = False
             update_dict["enable_db_preretriever"] = False
-        print("update_dict:", update_dict)
         rag_client.patch_config(update_dict)
     except RagApiError as api_error:
         raise gr.Error(f"HTTP {api_error.code} Error: {api_error.msg}")

diff --git a/src/pai_rag/app/web/tabs/settings_tab.py b/src/pai_rag/app/web/tabs/settings_tab.py
@@ -261,7 +261,7 @@ def create_setting_tab() -> Dict[str, Any]:
                     oss_endpoint = gr.Textbox(
                         label="OSS Endpoint",
                         elem_id="oss_endpoint",
-                        default="oss-cn-hangzhou.aliyuncs.com",
+                        placeholder="oss-cn-hangzhou.aliyuncs.com",
                     )
                 use_oss.input(
                     fn=ev_listeners.change_use_oss,

diff --git a/src/pai_rag/app/web/tabs/upload_tab.py b/src/pai_rag/app/web/tabs/upload_tab.py
@@ -124,7 +124,7 @@ def upload_knowledge(
 
     result = {"Info": ["StartTime", "EndTime", "Duration(s)", "Status"]}
     error_msg = ""
-    while not all(file.finished is True for file in my_upload_files):
+    while True:
         for file in my_upload_files:
             try:
                 response = asyncio.run(
@@ -144,12 +144,16 @@ def upload_knowledge(
             gr.update(visible=True, value=pd.DataFrame(result)),
             gr.update(visible=False),
         ]
-        if not all(file.finished is True for file in my_upload_files):
-            time.sleep(2)
+
+        if all(file.finished is True for file in my_upload_files):
+            break
+
+        time.sleep(2)
 
     upload_result = "Upload success."
     if error_msg:
         upload_result = f"Upload failed: {error_msg}"
+
     yield [
         gr.update(visible=True, value=pd.DataFrame(result)),
         gr.update(

diff --git a/src/pai_rag/app/web/tabs/vector_db_panel.py b/src/pai_rag/app/web/tabs/vector_db_panel.py
@@ -47,6 +47,7 @@ def create_vector_db_panel() -> Dict[str, Any]:
                         "cn-shenzhen",
                         "cn-chengdu",
                     ],
+                    value="cn-hangzhou",
                     label="RegionId",
                     elem_id="adb_region_id",
                 )

diff --git a/src/pai_rag/app/web/ui_constants.py b/src/pai_rag/app/web/ui_constants.py
@@ -19,16 +19,27 @@
 #             """
 
 WELCOME_MESSAGE = """
-            # \N{fire} PAI-RAG Dashboard
-
-            #### \N{rocket} Join the DingTalk Q&A Group: 27370042974
-            """
+        <div class="title"> <span class="head"> \N{fire} PAI-RAG Dashboard</span><span class="subhead"> \N{rocket} Join the DingTalk Q&A Group: 27370042974</span> </div>
+    """
 
 DEFAULT_CSS_STYPE = """
-        h1, h3, h4 {
+        .title {
+            margin-block-start: 0.2em;
+            margin-block-end: 0.2em;
             text-align: center;
             display:block;
         }
+
+        .head {
+            font-size: 2em;
+            font-weight: bold;
+        }
+
+        .subhead {
+            margin: 1.5em;
+            font-size: 1em;
+            font-weight: bold;
+        }
         """
 
 DEFAULT_EMBED_SIZE = 1024

diff --git a/src/pai_rag/app/web/webui.py b/src/pai_rag/app/web/webui.py
@@ -1,4 +1,3 @@
-from urllib.parse import urljoin
 from fastapi import FastAPI
 import gradio as gr
 from pai_rag.app.web import event_listeners
@@ -135,13 +134,11 @@ def configure_webapp(app: FastAPI, web_url, rag_url=DEFAULT_LOCAL_URL) -> gr.Blo
     rag_client.set_endpoint(rag_url)
 
     chat_page = create_chat_ui()
-    chat_page.queue(concurrency_count=1, max_size=64)
-    chat_page._queue.set_url(urljoin(web_url, "chat/"))
+    chat_page.queue(api_open=True, max_size=64)
     gr.mount_gradio_app(app, chat_page, path="/chat")
 
     home = make_homepage()
-    home.queue(concurrency_count=1, max_size=64)
-    home._queue.set_url(urljoin(web_url, "/"))
+    chat_page.queue(api_open=True, max_size=64)
     logger.info(f"web_url: {web_url}")
     gr.mount_gradio_app(app, home, path="/")
     return
diff --git a/src/pai_rag/evaluation/metrics/response/correctness.py b/src/pai_rag/evaluation/metrics/response/correctness.py
@@ -101,7 +101,6 @@ class Correctness(LlmMetric):
     Passing is defined as a score greater than or equal to the given threshold.
 
     Args:
-        service_context (Optional[ServiceContext]): Service context.
         eval_template (Optional[Union[BasePromptTemplate, str]]):
             Template for the evaluation prompt.
         score_threshold (float): Numerical threshold for passing the evaluation,

diff --git a/src/pai_rag/integrations/agent/pai/pai_agent.py b/src/pai_rag/integrations/agent/pai/pai_agent.py
@@ -7,7 +7,7 @@
     Type,
 )
 import json
-from llama_index.agent.openai.step import OpenAIAgentWorker
+from llama_index.agent.openai import OpenAIAgentWorker
 from llama_index.core.agent.runner.base import AgentRunner
 from llama_index.core.callbacks import CallbackManager
 from llama_index.core.base.llms.types import ChatMessage