bump llama_cloud and update pipeline

helpers/python.ts

@@ -103,7 +103,7 @@ const getAdditionalDependencies = (
     case "llamacloud":
       dependencies.push({
         name: "llama-index-indices-managed-llama-cloud",
-        version: "^0.6.0",
+        version: "^0.6.3",
       });
       break;
   }

templates/components/vectordbs/python/llamacloud/generate.py

@@ -1,68 +1,28 @@
 # flake8: noqa: E402
-import os
 
 from dotenv import load_dotenv
 
 load_dotenv()
 
 import logging
 
-from app.engine.index import get_client, get_index
+from llama_index.core.readers import SimpleDirectoryReader
+
+from app.engine.index import get_index
 from app.engine.service import LLamaCloudFileService  # type: ignore
 from app.settings import init_settings
-from llama_cloud import PipelineType
-from llama_index.core.readers import SimpleDirectoryReader
-from llama_index.core.settings import Settings
 
 logging.basicConfig(level=logging.INFO)
 logger = logging.getLogger()
 
 
-def ensure_index(index):
-    project_id = index._get_project_id()
-    client = get_client()
-    pipelines = client.pipelines.search_pipelines(
-        project_id=project_id,
-        pipeline_name=index.name,
-        pipeline_type=PipelineType.MANAGED.value,
-    )
-    if len(pipelines) == 0:
-        from llama_index.embeddings.openai import OpenAIEmbedding
-
-        if not isinstance(Settings.embed_model, OpenAIEmbedding):
-            raise ValueError(
-                "Creating a new pipeline with a non-OpenAI embedding model is not supported."
-            )
-        client.pipelines.upsert_pipeline(
-            project_id=project_id,
-            request={
-                "name": index.name,
-                "embedding_config": {
-                    "type": "OPENAI_EMBEDDING",
-                    "component": {
-                        "api_key": os.getenv("OPENAI_API_KEY"),  # editable
-                        "model_name": os.getenv("EMBEDDING_MODEL"),
-                    },
-                },
-                "transform_config": {
-                    "mode": "auto",
-                    "config": {
-                        "chunk_size": Settings.chunk_size,  # editable
-                        "chunk_overlap": Settings.chunk_overlap,  # editable
-                    },
-                },
-            },
-        )
-
-
 def generate_datasource():
     init_settings()
     logger.info("Generate index for the provided data")
 
-    index = get_index()
-    ensure_index(index)
-    project_id = index._get_project_id()
-    pipeline_id = index._get_pipeline_id()
+    index = get_index(create_if_missing=True)
+    if index is None:
+        raise ValueError("Index not found and could not be created")
 
     # use SimpleDirectoryReader to retrieve the files to process
     reader = SimpleDirectoryReader(
@@ -78,7 +38,7 @@ def generate_datasource():
                 f"Adding file {input_file} to pipeline {index.name} in project {index.project_name}"
             )
             LLamaCloudFileService.add_file_to_pipeline(
-                project_id, pipeline_id, f, custom_metadata={}
+                index.project.id, index.pipeline.id, f, custom_metadata={}
             )
 
     logger.info("Finished generating the index")

templates/components/vectordbs/python/llamacloud/index.py

@@ -2,10 +2,12 @@
 import os
 from typing import Optional
 
+from llama_cloud import PipelineType
 from llama_index.core.callbacks import CallbackManager
 from llama_index.core.ingestion.api_utils import (
     get_client as llama_cloud_get_client,
 )
+from llama_index.core.settings import Settings
 from llama_index.indices.managed.llama_cloud import LlamaCloudIndex
 from pydantic import BaseModel, Field, field_validator
 
@@ -82,14 +84,64 @@ def to_index_kwargs(self) -> dict:
         }
 
 
-def get_index(config: IndexConfig = None):
+def get_index(
+    config: IndexConfig = None,
+    create_if_missing: bool = False,
+):
     if config is None:
         config = IndexConfig()
-    index = LlamaCloudIndex(**config.to_index_kwargs())
-
-    return index
+    # Check whether the index exists
+    try:
+        index = LlamaCloudIndex(**config.to_index_kwargs())
+        return index
+    except ValueError:
+        if create_if_missing:
+            logger.info(
+                f"Index {config.llama_cloud_pipeline_config.pipeline} not found, creating it"
+            )
+            _create_index(config)
+            return LlamaCloudIndex(**config.to_index_kwargs())
+        return None
 
 
 def get_client():
     config = LlamaCloudConfig()
     return llama_cloud_get_client(**config.to_client_kwargs())
+
+
+def _create_index(
+    config: IndexConfig,
+):
+    client = get_client()
+    pipeline_name = config.llama_cloud_pipeline_config.pipeline
+
+    pipelines = client.pipelines.search_pipelines(
+        pipeline_name=pipeline_name,
+        pipeline_type=PipelineType.MANAGED.value,
+    )
+    if len(pipelines) == 0:
+        from llama_index.embeddings.openai import OpenAIEmbedding
+
+        if not isinstance(Settings.embed_model, OpenAIEmbedding):
+            raise ValueError(
+                "Creating a new pipeline with a non-OpenAI embedding model is not supported."
+            )
+        client.pipelines.upsert_pipeline(
+            request={
+                "name": pipeline_name,
+                "embedding_config": {
+                    "type": "OPENAI_EMBEDDING",
+                    "component": {
+                        "api_key": os.getenv("OPENAI_API_KEY"),  # editable
+                        "model_name": os.getenv("EMBEDDING_MODEL"),
+                    },
+                },
+                "transform_config": {
+                    "mode": "auto",
+                    "config": {
+                        "chunk_size": Settings.chunk_size,  # editable
+                        "chunk_overlap": Settings.chunk_overlap,  # editable
+                    },
+                },
+            },
+        )

templates/components/vectordbs/python/llamacloud/service.py

@@ -1,18 +1,18 @@
-from io import BytesIO
 import logging
 import os
 import time
-from typing import Any, Dict, List, Optional, Set, Tuple, Union
 import typing
+from io import BytesIO
+from typing import Any, Dict, List, Optional, Set, Tuple, Union
 
+import requests
 from fastapi import BackgroundTasks
 from llama_cloud import ManagedIngestionStatus, PipelineFileCreateCustomMetadataValue
+from llama_index.core.schema import NodeWithScore
 from pydantic import BaseModel
-import requests
+
 from app.api.routers.models import SourceNodes
 from app.engine.index import get_client
-from llama_index.core.schema import NodeWithScore
-
 
 logger = logging.getLogger("uvicorn")
 

templates/types/streaming/fastapi/pyproject.toml

@@ -19,7 +19,7 @@ python-dotenv = "^1.0.0"
 pydantic = "<2.10"
 aiostream = "^0.5.2"
 cachetools = "^5.3.3"
-llama-index = "^0.11.17"
+llama-index = "^0.12.1"
 rich = "^13.9.4"
 
 [tool.poetry.group.dev.dependencies]