fix: Reference is shown multiple times for every sentence in response

code/backend/batch/utilities/parser/output_parser_tool.py

@@ -20,17 +20,11 @@ def _get_source_docs_from_answer(self, answer):
         results = re.findall(r"\[doc(\d+)\]", answer)
         return [int(i) for i in results]
 
-    def _replace_last(self, text, old, new):
-        """Replaces the last occurence of a substring in a string
-
-        This is done by reversing the string using [::-1], replacing the first occurence of the reversed substring, and
-        reversing the string again.
-        """
-        return (text[::-1].replace(old[::-1], new[::-1], 1))[::-1]
-
-    def _make_doc_references_sequential(self, answer, doc_ids):
-        for i, idx in enumerate(doc_ids):
-            answer = self._replace_last(answer, f"[doc{idx}]", f"[doc{i+1}]")
+    def _make_doc_references_sequential(self, answer):
+        doc_matches = list(re.finditer(r"\[doc\d+\]", answer))
+        for i, match in enumerate(doc_matches):
+            start, end = match.span()
+            answer = answer[:start] + f"[doc{i + 1}]" + answer[end:]
         return answer
 
     def parse(
@@ -42,7 +36,7 @@ def parse(
     ) -> List[dict]:
         answer = self._clean_up_answer(answer)
         doc_ids = self._get_source_docs_from_answer(answer)
-        answer = self._make_doc_references_sequential(answer, doc_ids)
+        answer = self._make_doc_references_sequential(answer)
 
         # create return message object
         messages = [

code/frontend/src/components/Answer/AnswerParser.tsx

@@ -11,7 +11,7 @@ let filteredCitations = [] as Citation[];
 
 // Define a function to check if a citation with the same Chunk_Id already exists in filteredCitations
 const isDuplicate = (citation: Citation,citationIndex:string) => {
-    return filteredCitations.some((c) => c.chunk_id === citation.chunk_id) ;
+    return filteredCitations.some((c) => c.chunk_id === citation.chunk_id && c.id === citation.id) ;
 };
 
 export function parseAnswer(answer: AskResponse): ParsedAnswer {
@@ -28,12 +28,11 @@ export function parseAnswer(answer: AskResponse): ParsedAnswer {
         let citation = cloneDeep(answer.citations[Number(citationIndex) - 1]) as Citation;
         if (!isDuplicate(citation, citationIndex) && citation !== undefined) {
           answerText = answerText.replaceAll(link, ` ^${++citationReindex}^ `);
-          citation.id = citationIndex; // original doc index to de-dupe
           citation.reindex_id = citationReindex.toString(); // reindex from 1 for display
           filteredCitations.push(citation);
         }else{
             // Replacing duplicate citation with original index
-            let matchingCitation = filteredCitations.find((ct) => citation.chunk_id == ct.chunk_id);
+            let matchingCitation = filteredCitations.find((ct) => citation.chunk_id === ct.chunk_id && citation.id === ct.id);
             if (matchingCitation) {
                 answerText= answerText.replaceAll(link, ` ^${matchingCitation.reindex_id}^ `)
             }

infra/main.bicep

@@ -141,7 +141,7 @@ param azureOpenAIVisionModelVersion string = 'vision-preview'
 @description('Azure OpenAI Vision Model Capacity - See here for more info  https://learn.microsoft.com/en-us/azure/ai-services/openai/how-to/quota')
 param azureOpenAIVisionModelCapacity int = 10
 
-@description('Orchestration strategy: openai_function or semantic_kernel or langchain str. If you use a old version of turbo (0301), please select langchain')
+@description('Orchestration strategy: openai_function or semantic_kernel or langchain str. If you use a old version of turbo (0301), please select langchain. If the database type is PostgreSQL, set this to sementic_kernel.')
 @allowed([
   'openai_function'
   'semantic_kernel'
@@ -150,7 +150,7 @@ param azureOpenAIVisionModelCapacity int = 10
 ])
 param orchestrationStrategy string = 'semantic_kernel'
 
-@description('Chat conversation type: custom or byod.')
+@description('Chat conversation type: custom or byod. If the database type is PostgreSQL, set this to custom.')
 @allowed([
   'custom'
   'byod'
@@ -321,7 +321,7 @@ var eventGridSystemTopicName = 'doc-processing'
 var tags = { 'azd-env-name': environmentName }
 var rgName = 'rg-${environmentName}'
 var keyVaultName = 'kv-${resourceToken}'
-var baseUrl = 'https://raw.githubusercontent.com/Fr4nc3/chat-with-your-data-solution-accelerator/main/'
+var baseUrl = 'https://raw.githubusercontent.com/Azure-Samples/chat-with-your-data-solution-accelerator/main/'
 var azureOpenAIModelInfo = string({
   model: azureOpenAIModel
   modelName: azureOpenAIModelName
@@ -333,8 +333,8 @@ var azureOpenAIEmbeddingModelInfo = string({
   modelVersion: azureOpenAIEmbeddingModelVersion
 })
 
-var appversion = 'devpostgre' // Update GIT deployment branch
-var registryName = 'cwydcontainerregpk' // Update Registry name
+var appversion = 'latest' // Update GIT deployment branch
+var registryName = 'fruoccopublic' // Update Registry name
 
 // Organize resources in a resource group
 resource rg 'Microsoft.Resources/resourceGroups@2021-04-01' = {
@@ -383,7 +383,9 @@ module keyvault './core/security/keyvault.bicep' = if (useKeyVault || authType =
     location: location
     tags: tags
     principalId: principalId
-    managedIdentityObjectId: databaseType == 'PostgreSQL' ? managedIdentityModule.outputs.managedIdentityOutput.objectId : ''
+    managedIdentityObjectId: databaseType == 'PostgreSQL'
+      ? managedIdentityModule.outputs.managedIdentityOutput.objectId
+      : ''
   }
 }
 
@@ -877,13 +879,15 @@ module adminweb './app/adminweb.bicep' = if (hostingModel == 'code') {
         LOGLEVEL: logLevel
         DATABASE_TYPE: databaseType
       },
-      databaseType == 'PostgreSQL' ? {
-        AZURE_POSTGRESQL_INFO: string({
-          host: postgresDBModule.outputs.postgresDbOutput.postgreSQLServerName
-          dbname: postgresDBModule.outputs.postgresDbOutput.postgreSQLDatabaseName
-          user: adminWebsiteName
-        })
-      } : {}
+      databaseType == 'PostgreSQL'
+        ? {
+            AZURE_POSTGRESQL_INFO: string({
+              host: postgresDBModule.outputs.postgresDbOutput.postgreSQLServerName
+              dbname: postgresDBModule.outputs.postgresDbOutput.postgreSQLDatabaseName
+              user: adminWebsiteName
+            })
+          }
+        : {}
     )
   }
 }
@@ -961,13 +965,15 @@ module adminweb_docker './app/adminweb.bicep' = if (hostingModel == 'container')
         LOGLEVEL: logLevel
         DATABASE_TYPE: databaseType
       },
-      databaseType == 'PostgreSQL' ? {
-        AZURE_POSTGRESQL_INFO: string({
-          host: postgresDBModule.outputs.postgresDbOutput.postgreSQLServerName
-          dbname: postgresDBModule.outputs.postgresDbOutput.postgreSQLDatabaseName
-          user: '${adminWebsiteName}-docker'
-        })
-      } : {}
+      databaseType == 'PostgreSQL'
+        ? {
+            AZURE_POSTGRESQL_INFO: string({
+              host: postgresDBModule.outputs.postgresDbOutput.postgreSQLServerName
+              dbname: postgresDBModule.outputs.postgresDbOutput.postgreSQLDatabaseName
+              user: '${adminWebsiteName}-docker'
+            })
+          }
+        : {}
     )
   }
 }
@@ -1067,13 +1073,15 @@ module function './app/function.bicep' = if (hostingModel == 'code') {
         AZURE_SEARCH_TOP_K: azureSearchTopK
         DATABASE_TYPE: databaseType
       },
-      databaseType == 'PostgreSQL' ? {
-        AZURE_POSTGRESQL_INFO: string({
-            host: postgresDBModule.outputs.postgresDbOutput.postgreSQLServerName
-            dbname: postgresDBModule.outputs.postgresDbOutput.postgreSQLDatabaseName
-            user: functionName
-          })
-        } : {}
+      databaseType == 'PostgreSQL'
+        ? {
+            AZURE_POSTGRESQL_INFO: string({
+              host: postgresDBModule.outputs.postgresDbOutput.postgreSQLServerName
+              dbname: postgresDBModule.outputs.postgresDbOutput.postgreSQLDatabaseName
+              user: functionName
+            })
+          }
+        : {}
     )
   }
 }
@@ -1138,13 +1146,15 @@ module function_docker './app/function.bicep' = if (hostingModel == 'container')
         AZURE_SEARCH_TOP_K: azureSearchTopK
         DATABASE_TYPE: databaseType
       },
-      databaseType == 'PostgreSQL' ? {
-        AZURE_POSTGRESQL_INFO: string({
-            host: postgresDBModule.outputs.postgresDbOutput.postgreSQLServerName
-            dbname: postgresDBModule.outputs.postgresDbOutput.postgreSQLDatabaseName
-            user: '${functionName}-docker'
-          })
-        } : {}
+      databaseType == 'PostgreSQL'
+        ? {
+            AZURE_POSTGRESQL_INFO: string({
+              host: postgresDBModule.outputs.postgresDbOutput.postgreSQLServerName
+              dbname: postgresDBModule.outputs.postgresDbOutput.postgreSQLDatabaseName
+              user: '${functionName}-docker'
+            })
+          }
+        : {}
     )
   }
 }

infra/main.json

@@ -5,7 +5,7 @@
     "_generator": {
       "name": "bicep",
       "version": "0.30.23.60470",
-      "templateHash": "8207330457159201967"
+      "templateHash": "15536398457562727557"
     }
   },
   "parameters": {
@@ -312,7 +312,7 @@
         "prompt_flow"
       ],
       "metadata": {
-        "description": "Orchestration strategy: openai_function or semantic_kernel or langchain str. If you use a old version of turbo (0301), please select langchain"
+        "description": "Orchestration strategy: openai_function or semantic_kernel or langchain str. If you use a old version of turbo (0301), please select langchain. If the database type is PostgreSQL, set this to sementic_kernel."
       }
     },
     "conversationFlow": {
@@ -323,7 +323,7 @@
         "byod"
       ],
       "metadata": {
-        "description": "Chat conversation type: custom or byod."
+        "description": "Chat conversation type: custom or byod. If the database type is PostgreSQL, set this to custom."
       }
     },
     "azureOpenAITemperature": {
@@ -648,11 +648,11 @@
     },
     "rgName": "[format('rg-{0}', parameters('environmentName'))]",
     "keyVaultName": "[format('kv-{0}', parameters('resourceToken'))]",
-    "baseUrl": "https://raw.githubusercontent.com/Fr4nc3/chat-with-your-data-solution-accelerator/main/",
+    "baseUrl": "https://raw.githubusercontent.com/Azure-Samples/chat-with-your-data-solution-accelerator/main/",
     "azureOpenAIModelInfo": "[string(createObject('model', parameters('azureOpenAIModel'), 'modelName', parameters('azureOpenAIModelName'), 'modelVersion', parameters('azureOpenAIModelVersion')))]",
     "azureOpenAIEmbeddingModelInfo": "[string(createObject('model', parameters('azureOpenAIEmbeddingModel'), 'modelName', parameters('azureOpenAIEmbeddingModelName'), 'modelVersion', parameters('azureOpenAIEmbeddingModelVersion')))]",
-    "appversion": "devpostgre",
-    "registryName": "cwydcontainerregpk",
+    "appversion": "latest",
+    "registryName": "fruoccopublic",
     "defaultOpenAiDeployments": [
       {
         "name": "[parameters('azureOpenAIModel')]",