V.0.3.1.8 🐛 Bugfixes (#37)

* Updated Packages

* Bugfix: Check if requests exists before deleting

* Bugfix: Update Collections not working

Changed Put to Post req, moved to sub api route

backend/backend/app/api/routers/chat.py

@@ -4,10 +4,10 @@ from typing import List
 from fastapi import APIRouter, Depends, HTTPException, Request, status
 from fastapi.responses import StreamingResponse
 from fastapi.websockets import WebSocketDisconnect
-from llama_index.llms.base import ChatMessage
-from llama_index.llms.types import MessageRole
-from llama_index.memory import ChatMemoryBuffer
-from llama_index.prompts import PromptTemplate
+from llama_index.core.llms import ChatMessage
+from llama_index.core.llms import MessageRole
+from llama_index.core.memory import ChatMemoryBuffer
+from llama_index.core import PromptTemplate
 from pydantic import BaseModel
 
 from backend.app.utils import auth

backend/backend/app/api/routers/query.py

@@ -4,7 +4,7 @@ from typing import List
 from fastapi import APIRouter, Depends, HTTPException, Request, status
 from fastapi.responses import StreamingResponse
 from fastapi.websockets import WebSocketDisconnect
-from llama_index.llms.types import MessageRole
+from llama_index.core.llms import MessageRole
 from pydantic import BaseModel
 
 from backend.app.utils import auth

backend/backend/app/api/routers/search.py

@@ -2,8 +2,8 @@ import logging
 import re
 
 from fastapi import APIRouter, Depends, HTTPException, status
-from llama_index.postprocessor import SimilarityPostprocessor
-from llama_index.retrievers import VectorIndexRetriever
+from llama_index.core.postprocessor import SimilarityPostprocessor
+from llama_index.core.retrievers import VectorIndexRetriever
 
 from backend.app.utils import auth
 from backend.app.utils.index import get_index

backend/backend/app/utils/contants.py

@@ -29,7 +29,7 @@ LLM_TEMPERATURE = 0.1
 MODEL_KWARGS = {"n_gpu_layers": 100} if DEVICE_TYPE == "cuda" else {}
 
 # Service Context Constants
-CHUNK_SIZE = 1000
+CHUNK_SIZE = 1024
 CHUNK_OVERLAP = 100
 
 # Embedding Model Constants
@@ -37,12 +37,12 @@ EMBED_MODEL_NAME = "sentence-transformers/all-MiniLM-L6-v2"
 EMBED_POOLING = "mean"
 EMBED_MODEL_DIMENSIONS = 384  # MiniLM-L6-v2 uses 384 dimensions
 DEF_EMBED_MODEL_DIMENSIONS = (
-    1536  # Default embedding model dimensions used by OpenAI text-embedding-ada-002
+    1536  # Default embedding model dimensions used by OpenAI text-embedding-3-small
 )
 EMBED_BATCH_SIZE = 64  # batch size for openai embeddings
 
 # Chat Memory Buffer Constants
-MEMORY_TOKEN_LIMIT = 1500 if USE_LOCAL_LLM else 6144
+MEMORY_TOKEN_LIMIT = 3072 if USE_LOCAL_LLM else 6144
 
 # Prompt Helper Constants
 # set maximum input size

backend/backend/app/utils/index.py

@@ -2,23 +2,23 @@ import logging
 import os
 
 from dotenv import load_dotenv
-from llama_index import (
+from llama_index.core import (
     PromptHelper,
     ServiceContext,
-    # Document,
     SimpleDirectoryReader,
     StorageContext,
     VectorStoreIndex,
     load_index_from_storage,
     set_global_service_context,
 )
-from llama_index.embeddings import HuggingFaceEmbedding
-from llama_index.embeddings.openai import OpenAIEmbedding
-from llama_index.llms import LlamaCPP, OpenAI
-from llama_index.llms.llama_utils import (
+from llama_index.embeddings.openai import OpenAIEmbedding, OpenAIEmbeddingModelType
+from llama_index.legacy.embeddings.huggingface import HuggingFaceEmbedding
+from llama_index.legacy.llms.llama_utils import (
     completion_to_prompt,
     messages_to_prompt,
 )
+from llama_index.llms.llama_cpp import LlamaCPP
+from llama_index.llms.openai import OpenAI
 from llama_index.vector_stores.supabase import SupabaseVectorStore
 from vecs import IndexMeasure
 
@@ -45,7 +45,6 @@ from backend.app.utils.contants import (
     USE_LOCAL_VECTOR_STORE,
 )
 
-# from llama_index.vector_stores.supabase import SupabaseVectorStore
 # import textwrap
 
 load_dotenv()
@@ -98,7 +97,11 @@ else:
         api_key=os.getenv("OPENAI_API_KEY"),
     )
     # By default, LlamaIndex uses text-embedding-ada-002 from OpenAI
-    embed_model = OpenAIEmbedding(embed_batch_size=EMBED_BATCH_SIZE)
+    # Set the model to text-embed-3-small for better performance and cheaper cost
+    embed_model = OpenAIEmbedding(
+        model=OpenAIEmbeddingModelType.TEXT_EMBED_3_SMALL,
+        embed_batch_size=EMBED_BATCH_SIZE,
+    )
 
     prompt_helper = PromptHelper(
         chunk_size_limit=CHUNK_SIZE_LIMIT,
@@ -149,11 +152,11 @@ def create_index():
                     show_progress=True,
                 )
                 # store it for later
-                index.storage_context.persist(STORAGE_DIR)
-                logger.info(f"Finished creating new index. Stored in {STORAGE_DIR}")
+                index.storage_context.persist(new_storage_dir)
+                logger.info(f"Finished creating new index. Stored in {new_storage_dir}")
             else:
                 # do nothing
-                logger.info(f"Index already exist at {STORAGE_DIR}...")
+                logger.info(f"Index already exist at {new_storage_dir}...")
     # else, create & store the index in Supabase pgvector
     else:
         # get the folders in the data directory

backend/backend/app/utils/prompt_template.py

@@ -4,7 +4,7 @@ Modify the prompt template based on the model you select.
 This seems to have significant impact on the output of the LLM.
 """
 
-from llama_index.prompts import PromptTemplate
+from llama_index.core import PromptTemplate
 
 # this is specific to Llama-2.
 

backend/pyproject.toml

@@ -10,8 +10,7 @@ packages = [{ include = "backend" }]
 python = "^3.11,<3.12"
 fastapi = "^0.109.1"
 uvicorn = { extras = ["standard"], version = "^0.23.2" }
-llama-index = "^0.9.48"
-pypdf = "^3.17.4"
+pypdf = "^4.3.0"
 python-dotenv = "^1.0.0"
 llama-cpp-python = "^0.2.52"
 transformers = "^4.38.1"
@@ -22,6 +21,9 @@ pyjwt = "^2.8.0"
 vecs = "^0.4.3"
 python-multipart = "^0.0.9"
 asyncpg = "^0.29.0"
+llama-index = "^0.10.55"
+llama-index-vector-stores-supabase = "^0.1.5"
+llama-index-llms-llama-cpp = "^0.1.4"
 
 [tool.poetry.group.dev]
 optional = true

frontend/app/api/admin/collections/route.ts

@@ -38,44 +38,3 @@ export async function GET(request: NextRequest) {
 
     return NextResponse.json({ collections: collections });
 }
-
-// PUT request to update the collection data in the database
-export async function PUT(request: NextRequest) {
-    // Create a new Supabase client
-    const supabase = createClient(
-        process.env.SUPABASE_URL ?? '',
-        process.env.SUPABASE_SERVICE_ROLE_KEY ?? '',
-        { db: { schema: 'public' } },
-    );
-
-    // Retrieve the collection ID from the request body
-    const { collection_id, is_public } = await request.json();
-
-    // Update the collection data in the database
-    const { data: updateData, error: updateError } = await supabase
-        .from('collections')
-        .update({ is_public: is_public })
-        .eq('collection_id', collection_id);
-
-    if (updateError) {
-        console.error('Error updating collection data in database:', updateError.message);
-        return NextResponse.json({ error: updateError.message }, { status: 500 });
-    }
-
-    // console.log('Updated collection:', data);
-
-    // Delete the collection requests data in the database (Since it is manually updated by Admin)
-    const { data: delData, error: delError } = await supabase
-        .from('collections_requests')
-        .delete()
-        .eq('collection_id', collection_id);
-
-    if (delError) {
-        console.error('Error deleting collection requests data in database:', delError.message);
-        return NextResponse.json({ error: delError.message }, { status: 500 });
-    }
-
-    // console.log('Deleted collection requests:', delData);
-
-    return NextResponse.json({ message: 'Collection updated successfully' });
-}

frontend/app/api/admin/collections/update/route.ts

@@ -0,0 +1,60 @@
+export const revalidate = 10;
+
+import { createClient } from '@supabase/supabase-js';
+import { NextRequest, NextResponse } from "next/server";
+
+// POST request to update the collection data in the database
+export async function POST(request: NextRequest) {
+    // Create a new Supabase client
+    const supabase = createClient(
+        process.env.SUPABASE_URL ?? '',
+        process.env.SUPABASE_SERVICE_ROLE_KEY ?? '',
+        { db: { schema: 'public' } },
+    );
+
+    // Retrieve the collection ID from the request body
+    const { collection_id, is_public } = await request.json();
+
+    // Update the collection data in the database
+    const { data: updateData, error: updateError } = await supabase
+        .from('collections')
+        .update({ is_public: is_public })
+        .eq('collection_id', collection_id);
+
+    if (updateError) {
+        console.error('Error updating collection data in database:', updateError.message);
+        return NextResponse.json({ error: updateError.message }, { status: 500 });
+    }
+
+    // console.log('Updated collection:', updateData);
+
+    // Check if there is an existing collection request for the collection
+    const { data: collReq, error: collReqError } = await supabase
+        .from('collections_requests')
+        .select('collection_id')
+        .eq('collection_id', collection_id);
+
+    if (collReqError) {
+        console.error('Error fetching collection requests data from database:', collReqError.message);
+        return NextResponse.json({ error: collReqError.message }, { status: 500 });
+    }
+
+    // console.log('Collection requests:', collReq);
+
+    // If there is an existing collection request, delete it
+    if (collReq.length === 1) {
+        const { data: delData, error: delError } = await supabase
+            .from('collections_requests')
+            .delete()
+            .eq('collection_id', collection_id);
+
+        if (delError) {
+            console.error('Error deleting collection requests data in database:', delError.message);
+            return NextResponse.json({ error: delError.message }, { status: 500 });
+        }
+
+        // console.log('Deleted collection requests:', delData);
+    }
+
+    return NextResponse.json({ message: 'Collection updated successfully' });
+}

frontend/app/components/ui/admin/admin-manage-collections.tsx

@@ -70,8 +70,8 @@ export default function AdminManageCollections() {
         }).then((result) => {
             if (result.isConfirmed) {
                 // if user confirms, send request to server
-                fetch(`/api/admin/collections`, {
-                    method: 'PUT',
+                fetch(`/api/admin/collections/update`, {
+                    method: 'POST',
                     headers: {
                         'Content-Type': 'application/json',
                     },
@@ -131,8 +131,8 @@ export default function AdminManageCollections() {
         }).then((result) => {
             if (result.isConfirmed) {
                 // if user confirms, send request to server
-                fetch(`/api/admin/collections`, {
-                    method: 'PUT',
+                fetch(`/api/admin/collections/update`, {
+                    method: 'POST',
                     headers: {
                         'Content-Type': 'application/json',
                     },

frontend/package-lock.json

@@ -1,12 +1,12 @@
 {
   "name": "smart-retrieval",
-  "version": "0.1.0",
+  "version": "0.3.0",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "smart-retrieval",
-      "version": "0.1.0",
+      "version": "0.3.0",
       "dependencies": {
         "@auth/supabase-adapter": "^0.6.0",
         "@nextui-org/react": "^2.2.9",
@@ -4232,11 +4232,11 @@
       }
     },
     "node_modules/braces": {
-      "version": "3.0.2",
-      "resolved": "https://registry.npmjs.org/braces/-/braces-3.0.2.tgz",
-      "integrity": "sha512-b8um+L1RzM3WDSzvhm6gIz1yfTbBt6YTlcEKAvsmqCZZFw46z626lVj9j1yEPW33H5H+lBQpZMP1k8l+78Ha0A==",
+      "version": "3.0.3",
+      "resolved": "https://registry.npmjs.org/braces/-/braces-3.0.3.tgz",
+      "integrity": "sha512-yQbXgO/OSZVD2IsiLlro+7Hf6Q18EJrKSEsdoMzKePKXct3gvD8oLcOQdIzGupr5Fj+EDe8gO/lxc1BzfMpxvA==",
       "dependencies": {
-        "fill-range": "^7.0.1"
+        "fill-range": "^7.1.1"
       },
       "engines": {
         "node": ">=8"
@@ -5688,9 +5688,9 @@
       "peer": true
     },
     "node_modules/fast-loops": {
-      "version": "1.1.3",
-      "resolved": "https://registry.npmjs.org/fast-loops/-/fast-loops-1.1.3.tgz",
-      "integrity": "sha512-8EZzEP0eKkEEVX+drtd9mtuQ+/QrlfW/5MlwcwK5Nds6EkZ/tRzEexkzUY2mIssnAyVLT+TKHuRXmFNNXYUd6g=="
+      "version": "1.1.4",
+      "resolved": "https://registry.npmjs.org/fast-loops/-/fast-loops-1.1.4.tgz",
+      "integrity": "sha512-8dbd3XWoKCTms18ize6JmQF1SFnnfj5s0B7rRry22EofgMu7B6LKHVh+XfFqFGsqnbH54xgeO83PzpKI+ODhlg=="
     },
     "node_modules/fast-shallow-equal": {
       "version": "1.0.0",
@@ -5735,9 +5735,9 @@
       }
     },
     "node_modules/fill-range": {
-      "version": "7.0.1",
-      "resolved": "https://registry.npmjs.org/fill-range/-/fill-range-7.0.1.tgz",
-      "integrity": "sha512-qOo9F+dMUmC2Lcb4BbVvnKJxTPjCm+RRpe4gDuGrzkL7mEVl/djYSu2OdQ2Pa302N4oqkSg9ir6jaLWJ2USVpQ==",
+      "version": "7.1.1",
+      "resolved": "https://registry.npmjs.org/fill-range/-/fill-range-7.1.1.tgz",
+      "integrity": "sha512-YsGpe3WHLK8ZYi4tWDg2Jy3ebRz2rXowDxnld4bkQB00cc/1Zw9AWnC0i9ztDJitivtQvaI9KaLyKrc+hBW0yg==",
       "dependencies": {
         "to-regex-range": "^5.0.1"
       },
@@ -12758,9 +12758,9 @@
       "peer": true
     },
     "node_modules/sweetalert2": {
-      "version": "11.11.0",
-      "resolved": "https://registry.npmjs.org/sweetalert2/-/sweetalert2-11.11.0.tgz",
-      "integrity": "sha512-wKCTtoE6lQVDKaJ5FFq+znk/YykJmJlD8RnLZps8C7DyivctCoRlVeeOwnKfgwKS+QJYon7s++3dmNi3/am1tw==",
+      "version": "11.12.2",
+      "resolved": "https://registry.npmjs.org/sweetalert2/-/sweetalert2-11.12.2.tgz",
+      "integrity": "sha512-Rwv5iRYlApkDSXeX22aLhhWMlWPzFxnNBVLZajkFKYhaVEfQkMOPQQRhBtSFxKBPCoko9U3SccWm9hI4o3Id0Q==",
       "funding": {
         "type": "individual",
         "url": "https://github.com/sponsors/limonte"
@@ -13691,9 +13691,9 @@
       "peer": true
     },
     "node_modules/ws": {
-      "version": "8.16.0",
-      "resolved": "https://registry.npmjs.org/ws/-/ws-8.16.0.tgz",
-      "integrity": "sha512-HS0c//TP7Ina87TfiPUz1rQzMhHrl/SG2guqRcTOIUYD2q8uhUdNHZYJUaQ8aTGPzCh+c6oawMKW35nFl1dxyQ==",
+      "version": "8.18.0",
+      "resolved": "https://registry.npmjs.org/ws/-/ws-8.18.0.tgz",
+      "integrity": "sha512-8VbfWfHLbbwu3+N6OKsOMpBdT4kXPDDB9cJk2bJ6mh9ucxdlnNvH1e+roYkKmN9Nxw2yjz7VzeO9oOz2zJ04Pw==",
       "engines": {
         "node": ">=10.0.0"
       },