swap to new embedding model and handle user 'i dont know' scenario

config/model_config_advanced.yml

@@ -14,4 +14,4 @@ vector_store:
   persisted_path: './models/chroma_db_advanced'
 
 questionaire_data:
-  db_path: './database/mock_qna_advanced.sqlite'
+  db_path: './database/mock_qna.sqlite'

database/mock_qna.sqlite

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bae92f8657dd47a86cfb157a8b9a829115282746b04049a4d056a38af1b8ab4f
+oid sha256:2d51005d26f568ee304005ab7cf52cdc58a55f528230ae914a11dc9b75219623
 size 40960

models/chroma_db_advanced/af9795b7-8b5f-4493-adbc-40aedf3c96ed/data_level0.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:453d35bee81975816ce0a286e796c4884c609c148e52d0605ac221daa46bf3d7
+size 10056000

models/chroma_db_advanced/{a88943fe-4428-425d-8b9c-7bb8665a0c79 → af9795b7-8b5f-4493-adbc-40aedf3c96ed}/header.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e87a1dc8bcae6f2c4bea6d5dd5005454d4dace8637dae29bff3c037ea771411e
+oid sha256:89bd0cf182f20a10a0d7faa81bf3304c0565bc9b6f4705056ae63c061b9269ff
 size 100

models/chroma_db_advanced/{a88943fe-4428-425d-8b9c-7bb8665a0c79/data_level0.bin → af9795b7-8b5f-4493-adbc-40aedf3c96ed/index_metadata.pickle}

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2eec38a208011f4f233e59d2618152fa02e42d91757412778a5db814fe80bf2f
-size 1676000
+oid sha256:c5ecccac152d2deee938b41b1533b454bb8d5778a0befcd855529538a1a17bdf
+size 346049

models/chroma_db_advanced/{a88943fe-4428-425d-8b9c-7bb8665a0c79 → af9795b7-8b5f-4493-adbc-40aedf3c96ed}/length.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fc19b1997119425765295aeab72d76faa6927d4f83985d328c26f20468d6cc76
-size 4000
+oid sha256:ae1fb78e4b679db0ad051360ddb549f4584c14a8b45f99d8d052f7d67067acb3
+size 24000

models/chroma_db_advanced/af9795b7-8b5f-4493-adbc-40aedf3c96ed/link_lists.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:535d672bfbbeec1181b50015d78bc1e776088cbbb0738d04bc725a76249eb744
+size 52152

models/chroma_db_advanced/chroma.sqlite3

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:51aba6bb0bf5e5851de1e4e6cf53215b874c11b7194b3b765a2edfbc59ce9313
-size 15937536
+oid sha256:74c0d3543bf7cab83459feda7fad58a984a7c018fc566f79e937038b3756fcca
+size 101720064

notebooks/002_persisted-embedding-model-advanced.ipynb

@@ -10,11 +10,16 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 1,
    "id": "7de9c591-5a77-4bbe-80f1-4897e15f0b97",
    "metadata": {},
    "outputs": [],
    "source": [
+    "import sys, os, shutil\n",
+    "sys.path.insert(0, \"../\")\n",
+    "\n",
+    "from preprocess_raw_documents import split_content\n",
+    "\n",
     "import chromadb\n",
     "from llama_index.core import VectorStoreIndex, SimpleDirectoryReader\n",
     "from llama_index.vector_stores.chroma.base import ChromaVectorStore\n",
@@ -31,27 +36,78 @@
     "import time"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "978152ce-4d87-44b5-b521-dbaff60b32b0",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "199it [00:00, 8821.71it/s]\n",
+      "200it [00:00, 12584.17it/s]\n"
+     ]
+    }
+   ],
+   "source": [
+    "split_content(filepath=\"../raw_documents/answers.txt\", \n",
+    "              separator=\"\\n\\n\", \n",
+    "              tmp_folder=\"../raw_documents/answers_temp\")\n",
+    "\n",
+    "split_content(filepath=\"../raw_documents/qna.txt\", \n",
+    "              separator=\"\\n\\n\\n\", \n",
+    "              tmp_folder=\"../raw_documents/qna_temp\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "d925371b-8777-4f5b-a7f2-ec3f228ef266",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "answers_temp_files = []\n",
+    "folder_path = \"../raw_documents/answers_temp\"\n",
+    "for f in os.listdir(folder_path):\n",
+    "    fpath = os.path.join(folder_path, f)\n",
+    "    answers_temp_files.append(fpath)\n",
+    "    \n",
+    "qna_temp_files = []\n",
+    "folder_path = \"../raw_documents/qna_temp\"\n",
+    "for f in os.listdir(folder_path):\n",
+    "    fpath = os.path.join(folder_path, f)\n",
+    "    qna_temp_files.append(fpath)"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,
+   "id": "e876a26b-822d-44d6-a3dd-ccdcc04933cf",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
    "id": "3e65dff6-77b6-4be8-8857-5cecf3a035bb",
    "metadata": {},
    "outputs": [],
    "source": [
     "# load some documents\n",
     "documents = SimpleDirectoryReader(input_files=[\n",
-    "                                    \"../raw_documents/qna.txt\",\n",
     "                                    \"../raw_documents/HI Chapter Summary Version 1.3.pdf\",\n",
     "                                    \"../raw_documents/conversation_examples.txt\",\n",
     "                                    \"../raw_documents/HI_Knowledge_Base.pdf\",\n",
-    "                                    \"../raw_documents/answers.txt\",\n",
-    "                                  ]).load_data()\n",
+    "                                  ] + answers_temp_files + qna_temp_files ).load_data()\n",
     "document = Document(text=\"\\n\\n\".join([doc.text for doc in documents]))"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 8,
    "id": "bd86b3f5-1dfc-4257-bd9c-86d34f02398d",
    "metadata": {},
    "outputs": [],
@@ -62,7 +118,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 9,
    "id": "f568ce7b-bcbf-455c-acf1-6c2cae129fed",
    "metadata": {},
    "outputs": [],
@@ -73,7 +129,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 10,
    "id": "ed0b018e-1982-46b2-b1b4-04f5c0ce8672",
    "metadata": {},
    "outputs": [],
@@ -92,19 +148,28 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 11,
    "id": "0946b6ce-96ab-44de-ad75-e424a8429f67",
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "LLM is explicitly disabled. Using MockLLM.\n"
+     ]
+    }
+   ],
    "source": [
     "Settings.llm = None\n",
     "Settings.chunk_size = 1024\n",
+    "Settings.chunk_overlap = 50\n",
     "Settings.embed_model = \"local:../models/fine-tuned-embeddings-advanced\""
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 12,
    "id": "b8c73a2c-1129-406a-8046-085afcaf9cbb",
    "metadata": {},
    "outputs": [],
@@ -114,10 +179,21 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 13,
    "id": "75f1c76f-d3e5-4b69-818c-98865adb1457",
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "6814"
+      ]
+     },
+     "execution_count": 13,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
    "source": [
     "len(nodes)"
    ]
@@ -132,7 +208,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 14,
    "id": "dab4c6f3-ef67-4d90-b3d5-e290c5d1b6f4",
    "metadata": {},
    "outputs": [],
@@ -142,7 +218,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 15,
    "id": "6a764113-ad7e-4674-aa57-ebbf405902a8",
    "metadata": {},
    "outputs": [],
@@ -160,7 +236,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 16,
    "id": "e492ed4a-23a3-47d6-8b50-51fb48b3aa05",
    "metadata": {},
    "outputs": [],
@@ -170,7 +246,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 17,
    "id": "cbd11b89-9b83-4f08-bb30-160f750f2ffb",
    "metadata": {},
    "outputs": [],
@@ -180,39 +256,88 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
-   "id": "082a0d7e-b025-4db1-be2a-7a0b7bc453b9",
+   "execution_count": 18,
+   "id": "d3bd848d-9985-4a3d-bdc4-ec340cc69ef3",
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Indexing time: 2.3 mins\n"
+     ]
+    }
+   ],
    "source": [
-    "vector_query_engine = vector_index.as_query_engine()"
+    "indexing_cost = time.time() - start_time\n",
+    "indexing_cost = indexing_cost / 60\n",
+    "print(f\"Indexing time: {indexing_cost:.1f} mins\")"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": null,
-   "id": "d3bd848d-9985-4a3d-bdc4-ec340cc69ef3",
+   "execution_count": 19,
+   "id": "f16cca33-71fb-437d-a033-671b9fd44054",
    "metadata": {},
    "outputs": [],
    "source": [
-    "indexing_cost = time.time() - start_time\n",
-    "indexing_cost = indexing_cost / 60\n",
-    "print(f\"Indexing time: {indexing_cost:.1f} mins\")"
+    "vector_query_engine = vector_index.as_query_engine()"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 20,
    "id": "3290e870-41d7-49c4-9c4f-cb16bd1f469e",
    "metadata": {
     "scrolled": true
    },
-   "outputs": [],
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "Response(response='Context information is below.\\n---------------------\\nfile_path: ../raw_documents/answers_temp/answers_050.txt\\n\\nQuestion: The fundamental principle of Singapore healthcare financing is ____________.\\nAnswer: The answer is \"Individual Savings\".\\n\\nfile_path: ../raw_documents/qna_temp/qna_050.txt\\n\\nC1/5\\nQuestion: The fundamental principle of Singapore healthcare financing is ____________.\\nA. The 3’s M. That is Medisave, Medishield, Medifund.\\nB. Means Testing and Casemix.\\nC. Individual Savings.\\nD. Tax based subsidies and government subvention.\\nAnswer: C. The answer is \"Individual Savings\".\\n---------------------\\nGiven the context information and not prior knowledge, answer the query.\\nQuery: Healthcare System in Singapore consists of?\\nAnswer: ', source_nodes=[NodeWithScore(node=TextNode(id_='536fef67-6a3f-4054-a94a-cc9143599510', embedding=None, metadata={'file_path': '../raw_documents/answers_temp/answers_050.txt', 'file_name': 'answers_050.txt', 'file_type': 'text/plain', 'file_size': 130, 'creation_date': '2024-02-24', 'last_modified_date': '2024-02-24', 'last_accessed_date': '2024-02-24'}, excluded_embed_metadata_keys=['file_name', 'file_type', 'file_size', 'creation_date', 'last_modified_date', 'last_accessed_date'], excluded_llm_metadata_keys=['file_name', 'file_type', 'file_size', 'creation_date', 'last_modified_date', 'last_accessed_date'], relationships={<NodeRelationship.SOURCE: '1'>: RelatedNodeInfo(node_id='2b0f7dad-c532-4abd-8c42-f53383a4fc76', node_type=<ObjectType.DOCUMENT: '4'>, metadata={'file_path': '../raw_documents/answers_temp/answers_050.txt', 'file_name': 'answers_050.txt', 'file_type': 'text/plain', 'file_size': 130, 'creation_date': '2024-02-24', 'last_modified_date': '2024-02-24', 'last_accessed_date': '2024-02-24'}, hash='5b1d1dc729a663e4ccfacc0f18adf0f6644a2a7d2991490fd962d1550c83f2ff'), <NodeRelationship.PREVIOUS: '2'>: RelatedNodeInfo(node_id='6d93c092-b4cc-4b5b-b379-080d777d3908', node_type=<ObjectType.TEXT: '1'>, metadata={'file_path': '../raw_documents/answers_temp/answers_044.txt', 'file_name': 'answers_044.txt', 'file_type': 'text/plain', 'file_size': 164, 'creation_date': '2024-02-24', 'last_modified_date': '2024-02-24', 'last_accessed_date': '2024-02-24'}, hash='caeb59043b8daa56ed472941882947570abff951f64aa0498672aba5921fac1d'), <NodeRelationship.NEXT: '3'>: RelatedNodeInfo(node_id='859a9958-6f5d-4581-95d0-39edfc950ef5', node_type=<ObjectType.TEXT: '1'>, metadata={}, hash='8416454b2fbad3e6122c5151d2b3d1eadf0afde3514ba09374c71e96baf712bc')}, text='Question: The fundamental principle of Singapore healthcare financing is ____________.\\nAnswer: The answer is \"Individual Savings\".', start_char_idx=0, end_char_idx=130, text_template='{metadata_str}\\n\\n{content}', metadata_template='{key}: {value}', metadata_seperator='\\n'), score=0.4159636550867191), NodeWithScore(node=TextNode(id_='472000ae-a0aa-4464-a200-72fe67a3fbde', embedding=None, metadata={'file_path': '../raw_documents/qna_temp/qna_050.txt', 'file_name': 'qna_050.txt', 'file_type': 'text/plain', 'file_size': 297, 'creation_date': '2024-02-24', 'last_modified_date': '2024-02-24', 'last_accessed_date': '2024-02-24'}, excluded_embed_metadata_keys=['file_name', 'file_type', 'file_size', 'creation_date', 'last_modified_date', 'last_accessed_date'], excluded_llm_metadata_keys=['file_name', 'file_type', 'file_size', 'creation_date', 'last_modified_date', 'last_accessed_date'], relationships={<NodeRelationship.SOURCE: '1'>: RelatedNodeInfo(node_id='506fb715-d3b0-4ca7-b7ca-011a1e1a1f0d', node_type=<ObjectType.DOCUMENT: '4'>, metadata={'file_path': '../raw_documents/qna_temp/qna_050.txt', 'file_name': 'qna_050.txt', 'file_type': 'text/plain', 'file_size': 297, 'creation_date': '2024-02-24', 'last_modified_date': '2024-02-24', 'last_accessed_date': '2024-02-24'}, hash='7461ffa12ff6729003131976b82995b7254ab10f8dc7d79c65988ec9e3b7b631'), <NodeRelationship.PREVIOUS: '2'>: RelatedNodeInfo(node_id='d8232b90-d641-4966-b98f-4ca0821db773', node_type=<ObjectType.TEXT: '1'>, metadata={'file_path': '../raw_documents/qna_temp/qna_044.txt', 'file_name': 'qna_044.txt', 'file_type': 'text/plain', 'file_size': 383, 'creation_date': '2024-02-24', 'last_modified_date': '2024-02-24', 'last_accessed_date': '2024-02-24'}, hash='cbeb00c29c6130548466697a862fee43ab2be92d84158cc0b69c2f5c7bbe68b1'), <NodeRelationship.NEXT: '3'>: RelatedNodeInfo(node_id='e772e623-cf91-41cd-a516-50acb894eb54', node_type=<ObjectType.TEXT: '1'>, metadata={}, hash='a7583b0fd46f98d0118c712632277d81f417b779f8bcc100ab2558dae6317cde')}, text='C1/5\\nQuestion: The fundamental principle of Singapore healthcare financing is ____________.\\nA. The 3’s M. That is Medisave, Medishield, Medifund.\\nB. Means Testing and Casemix.\\nC. Individual Savings.\\nD. Tax based subsidies and government subvention.\\nAnswer: C. The answer is \"Individual Savings\".', start_char_idx=0, end_char_idx=295, text_template='{metadata_str}\\n\\n{content}', metadata_template='{key}: {value}', metadata_seperator='\\n'), score=0.4126648577998099)], metadata={'536fef67-6a3f-4054-a94a-cc9143599510': {'file_path': '../raw_documents/answers_temp/answers_050.txt', 'file_name': 'answers_050.txt', 'file_type': 'text/plain', 'file_size': 130, 'creation_date': '2024-02-24', 'last_modified_date': '2024-02-24', 'last_accessed_date': '2024-02-24'}, '472000ae-a0aa-4464-a200-72fe67a3fbde': {'file_path': '../raw_documents/qna_temp/qna_050.txt', 'file_name': 'qna_050.txt', 'file_type': 'text/plain', 'file_size': 297, 'creation_date': '2024-02-24', 'last_modified_date': '2024-02-24', 'last_accessed_date': '2024-02-24'}})"
+      ]
+     },
+     "execution_count": 20,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
    "source": [
     "response = vector_query_engine.query(\"Healthcare System in Singapore consists of?\")\n",
     "response"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "aa4b9906-5f75-4003-9f4c-5cfcc7ab1eaf",
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 21,
+   "id": "1bb75b04-6a62-43a4-8728-d2e52e49f1c0",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "if os.path.exists(\"../raw_documents/answers_temp\"):\n",
+    "    shutil.rmtree(\"../raw_documents/answers_temp\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 22,
+   "id": "0ed920fb-6456-49ac-8b63-08bd86b5b39c",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "if os.path.exists(\"../raw_documents/qna_temp\"):\n",
+    "    shutil.rmtree(\"../raw_documents/qna_temp\")"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,
@@ -239,7 +364,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 1,
    "id": "c1a42c35-5f57-423c-8fb7-7d18b3b466b5",
    "metadata": {},
    "outputs": [],
@@ -269,7 +394,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 2,
    "id": "d38dc953-b923-4128-86a1-c8c6f69af0ed",
    "metadata": {},
    "outputs": [],
@@ -279,7 +404,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 3,
    "id": "4c83c613-2cfc-4871-9d07-c82f77a3bd5e",
    "metadata": {},
    "outputs": [],
@@ -289,7 +414,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 4,
    "id": "0583e9b0-d977-488c-8331-46dfa749924c",
    "metadata": {},
    "outputs": [],
@@ -308,7 +433,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 5,
    "id": "2159a2b6-494b-41b9-ac54-dd342bfb74ba",
    "metadata": {},
    "outputs": [],
@@ -318,7 +443,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 6,
    "id": "1b385644-b46e-4d13-88fa-9f4af39db405",
    "metadata": {},
    "outputs": [],
@@ -328,7 +453,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 7,
    "id": "93cb53d1-6b8c-4b2d-a839-53501c0d54b2",
    "metadata": {},
    "outputs": [],
@@ -340,7 +465,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 8,
    "id": "c40d59e1-6d42-41f0-8c9b-70aa026093ae",
    "metadata": {},
    "outputs": [],
@@ -362,7 +487,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 9,
    "id": "1a506940-c2b4-4d14-ad93-fd451331c582",
    "metadata": {},
    "outputs": [],
@@ -375,7 +500,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 10,
    "id": "3f592848-8536-4b4d-b34a-adc32d043432",
    "metadata": {},
    "outputs": [],
@@ -385,7 +510,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 11,
    "id": "6c7df81a-fd2f-42bf-b09c-46d7750f7252",
    "metadata": {},
    "outputs": [],
@@ -399,58 +524,66 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
-   "id": "434f0caf-8b1f-40c6-b9ec-b039cd1ca612",
+   "execution_count": 12,
+   "id": "c3106dff-dd6f-47a9-9454-1e61775e7539",
    "metadata": {},
    "outputs": [],
    "source": [
-    "prompt = \"\"\"\n",
-    "Question: Which of the following is NOT a characteristic of medical expense insurance?\n",
-    "A. Pro ration factor and co-insurance.\n",
-    "B. Deductibles apply for all treatments.\n",
-    "C. Impose Sub- Limits.\n",
-    "D. Can be issued as a rider or stand-alone.\n",
-    "\"\"\""
+    "hi_engine = index.as_query_engine(\n",
+    "                memory=memory,\n",
+    "                system_prompt=system_content,\n",
+    "                similarity_top_k=10,\n",
+    "                streaming=True\n",
+    ")"
    ]
   },
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "78abaf95-e52d-445c-9d8e-bc51efb20f06",
+   "id": "53a38081-4a79-44bc-bfa3-5d8653804328",
    "metadata": {},
    "outputs": [],
-   "source": [
-    "res = chat_engine.chat(prompt)\n",
-    "print(res.response)"
-   ]
+   "source": []
   },
   {
    "cell_type": "code",
-   "execution_count": null,
-   "id": "1e62303c-3a00-448f-ad93-15cb6cee1f24",
+   "execution_count": 24,
+   "id": "434f0caf-8b1f-40c6-b9ec-b039cd1ca612",
    "metadata": {},
    "outputs": [],
-   "source": []
+   "source": [
+    "prompt = \"\"\"\n",
+    "Question: Which is not a government healthcare philosophy? \n",
+    "A. To nurture a healthy nation by promoting good health.\n",
+    "B. To rely on competition to improve service and raise efficiency\n",
+    "C. To intervene directly whenever necessary\n",
+    "D. To provide for the care of employees\n",
+    "\"\"\""
+   ]
   },
   {
    "cell_type": "code",
-   "execution_count": null,
-   "id": "dad72f9f-7f86-407d-93be-f5724cb30d5c",
+   "execution_count": 26,
+   "id": "a1c83dff-50d1-47b1-b7e9-4fc5cd08e1e8",
    "metadata": {},
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "D. To provide for the care of employees\n"
+     ]
+    }
+   ],
    "source": [
-    "hi_engine = index.as_query_engine(\n",
-    "                memory=memory,\n",
-    "                system_prompt=system_content,\n",
-    "                similarity_top_k=3,\n",
-    "                streaming=True\n",
-    ")"
+    "res = hi_engine.query(prompt)\n",
+    "print(res)"
    ]
   },
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "ab778a5d-d438-4f39-88f5-c67a1f1d575e",
+   "id": "cedd3512-548d-4455-80fd-c6a8b2c0cd00",
    "metadata": {},
    "outputs": [],
    "source": []
@@ -458,18 +591,44 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "7bb7c21a-7461-40c1-87a7-4a1f92f70153",
+   "id": "ec53dfcf-d4c0-4d10-a24e-be2004a83656",
    "metadata": {},
    "outputs": [],
+   "source": []
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "id": "78abaf95-e52d-445c-9d8e-bc51efb20f06",
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...\n",
+      "To disable this warning, you can either:\n",
+      "\t- Avoid using `tokenizers` before the fork if possible\n",
+      "\t- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "The correct answer is \"Deductibles apply for all treatments\".\n"
+     ]
+    }
+   ],
    "source": [
-    "res = hi_engine.query(\"may I know what is the rationale?\")\n",
-    "print(res)"
+    "res = chat_engine.chat(prompt)\n",
+    "print(res.response)"
    ]
   },
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "874a39ce-e682-42fa-8085-646bacea6cdb",
+   "id": "1e62303c-3a00-448f-ad93-15cb6cee1f24",
    "metadata": {},
    "outputs": [],
    "source": []

preprocess_raw_documents.py

@@ -0,0 +1,16 @@
+import os
+import shutil
+from tqdm import tqdm
+
+
+def split_content(filepath, separator, tmp_folder):
+    os.makedirs(tmp_folder, exist_ok=True)
+    base_file_name = os.path.basename(filepath)
+    fname, fextn = base_file_name.split(".")
+    with open(filepath, "r") as fp:
+        content = fp.read()
+    content_chunk = content.split(separator)
+    for index, chunk in tqdm(enumerate(content_chunk)):
+        new_fpath = os.path.join(tmp_folder, f"{fname}_{index:03d}.{fextn}")
+        with open(new_fpath, "w") as fp:
+            fp.write(chunk)

qna_prompting.py

@@ -25,10 +25,11 @@ qna_question_data_format = """
         Example 3: `Chapter_5` for fifth chapter
 """
 qna_answer_description = """
-    Use this tool to trigger the evaluation of user's provided input with the 
-    correct answer of the Q&A question asked. When user provides answer to the
-    question asked, they can reply in natural language or giving the alphabet
-    letter of which selected choice they think it's the right answer.
+    Not to trigger this when questions being asked, come directly from user. 
+    Only use this tool to trigger the evaluation of user's provided input with the 
+    correct answer of the Q&A question asked by Assistant. When user provides 
+    answer to the question asked, they can reply in natural language or giving 
+    the alphabet letter of which selected choice they think it's the right answer.
     
     If user's answer is not a single alphabet letter, but is contextually 
     closer to a particular answer choice, return the corresponding
@@ -122,7 +123,6 @@ def evaluate_qna_answer(user_selected_answer: str) -> str:
 
         ### convert to numeric type
         qna_answer  = int(qna_answer)
-        
         qna_answer_alphabet = num_mapping.get(qna_answer, "ERROR")
 
         con = sqlite3.connect(db_path)
@@ -138,13 +138,34 @@ def evaluate_qna_answer(user_selected_answer: str) -> str:
         con.commit()
         con.close()
 
-        if qna_answer == user_answer_numeric:
+        reasoning = "" if "textbook" in reasons else f"Rationale is that: {reasons}. "
+        qna_answer_response = (
+            f"Your selected answer is `{user_selected_answer}`, "
+            f"but the actual answer is `{qna_answer_alphabet}`. "
+        )
+        qna_not_knowing_response = (
+            f"No problem! The answer is `{qna_answer_alphabet}`. "
+        )
+        to_know_more = (
+            "Let me know if you want to know more, "
+            "I can give you an explanation 🐻💕"
+        )
+
+        if user_answer_numeric == 0:
+            st.toast("🍯❓ couldn't find the honey? 👌 no worries!", icon="🫠")
+            time.sleep(2)
+            st.toast("🐻 Let me bring it to you! 🍯💕", icon="💌")
+            time.sleep(2)
+            st.toast("✨ You will do great next time! 💆", icon="🎁")
+            final_response = qna_not_knowing_response + reasoning + to_know_more
+        elif qna_answer == user_answer_numeric:
             st.toast("🍯 yummy yummy, hooray!", icon="🎉")
             time.sleep(2)
             st.toast("🐻💕🍯 You got it right!", icon="🎊")
             time.sleep(2)
             st.toast("🥇 You are amazing! 💯💯", icon="💪")
             st.balloons()
+            final_response = qna_answer_response + reasoning + to_know_more
         else:
             st.toast("🐼 Something doesn't seem right.. 🔥🏠🔥", icon="😂")
             time.sleep(2)
@@ -152,17 +173,16 @@ def evaluate_qna_answer(user_selected_answer: str) -> str:
             time.sleep(2)
             st.toast("🤜🤛 Nevertheless, it was a good try!! 🏋️‍♂️🏋️‍♂️", icon="👏")
             st.snow()
+            final_response = qna_answer_response + reasoning + to_know_more
+
+        st.session_state.question_id = None
+        st.session_state.qna_answer = None
+        st.session_state.reasons = None
 
-        reasoning = "" if "textbook" in reasons else "Rationale is that: " + reasons
-        qna_answer_response = (
-            f"Your selected answer is `{user_selected_answer}`, "
-            f"but the actual answer is `{qna_answer_alphabet}`. " + reasoning
-        )
-        
     except Exception as e:
         print(e)
 
-    return qna_answer_response
+    return final_response
 
 get_qna_question_tool = FunctionTool.from_defaults(
                             fn=get_qna_question,

streamlit_app.py

@@ -40,7 +40,7 @@ nest_asyncio.apply()
 st.set_page_config(page_title="🐻📚 Study Bear 🍯")
 openai_api = os.getenv("OPENAI_API_KEY")
 
-with open("./config/model_config.yml", "r") as file_reader:
+with open("./config/model_config_advanced.yml", "r") as file_reader:
     model_config = yaml.safe_load(file_reader)
 
 input_files = model_config["input_data"]["source"]