Automatisation OpenAI avec n8n : génération de graphiques interactifs

{
  "meta": {
    "instanceId": "f4f5d195bb2162a0972f737368404b18be694648d365d6c6771d7b4909d28167",
    "templateCredsSetupCompleted": true
  },
  "nodes": [
    {
      "id": "50695e7f-3334-4124-a46e-1b3819412e26",
      "name": "OpenAI Chat Model",
      "type": "@n8n/n8n-nodes-langchain.lmChatOpenAi",
      "position": [
        1260,
        560
      ],
      "parameters": {
        "model": "gpt-4o",
        "options": {
          "temperature": 0.1
        }
      },
      "credentials": {
        "openAiApi": {
          "id": "WqzqjezKh8VtxdqA",
          "name": "OpenAi account - Baptiste"
        }
      },
      "typeVersion": 1
    },
    {
      "id": "2f07481d-3ca4-48ab-a8ff-59e9ab5c6062",
      "name": "Execute Workflow",
      "type": "n8n-nodes-base.executeWorkflow",
      "position": [
        2360,
        280
      ],
      "parameters": {
        "options": {
          "waitForSubWorkflow": true
        },
        "workflowId": {
          "__rl": true,
          "mode": "id",
          "value": "={{ $workflow.id }}"
        }
      },
      "typeVersion": 1.1
    },
    {
      "id": "49120164-4ffc-4fe0-8ee3-4ae13bda6c8d",
      "name": "Execute \"Generate a chart\" tool",
      "type": "n8n-nodes-base.executeWorkflowTrigger",
      "position": [
        1320,
        1140
      ],
      "parameters": {},
      "typeVersion": 1
    },
    {
      "id": "0fc6eaf9-8521-44ec-987e-73644d0cba79",
      "name": "OpenAI - Generate Chart definition with Structured Output",
      "type": "n8n-nodes-base.httpRequest",
      "position": [
        1620,
        1140
      ],
      "parameters": {
        "url": "https://api.openai.com/v1/chat/completions",
        "method": "POST",
        "options": {},
        "jsonBody": "={\n    \"model\": \"gpt-4o-2024-08-06\",\n    \"messages\": [\n        {\n            \"role\": \"system\",\n            \"content\": \"Based on the user request, generate a valid Chart.js definition. Important: - Be careful with the data scale and beginatzero that all data are visible. Example if ploted data 2 and 3 on a bar chart, the baseline should be 0. - Charts colors should be different only if there are multiple datasets. - Output valid JSON. In scales, min and max are numbers. Example: `{scales:{yAxes:[{ticks:{min:0,max:3}`\"\n        },\n        {\n            \"role\": \"user\",\n            \"content\": \"**User Request**: {{ $json.user_question }} \\n **Data to visualize**: {{ $json.output.replaceAll('\\n', \" \").replaceAll('\"', \"\") }}\"\n        }\n    ],\n    \"response_format\": {\n  \"type\": \"json_schema\",\n  \"json_schema\": {\n    \"name\": \"chart_configuration\",\n    \"description\": \"Configuration schema for Chart.js charts\",\n    \"strict\": true,\n    \"schema\": {\n  \"type\": \"object\",\n  \"properties\": {\n    \"type\": {\n      \"type\": \"string\",\n      \"enum\": [\"bar\", \"line\", \"radar\", \"pie\", \"doughnut\", \"polarArea\", \"bubble\", \"scatter\", \"area\"]\n    },\n    \"data\": {\n      \"type\": \"object\",\n      \"properties\": {\n        \"labels\": {\n          \"type\": \"array\",\n          \"items\": {\n            \"type\": \"string\"\n          }\n        },\n        \"datasets\": {\n          \"type\": \"array\",\n          \"items\": {\n            \"type\": \"object\",\n            \"properties\": {\n              \"label\": {\n                \"type\": [\"string\", \"null\"]\n              },\n              \"data\": {\n                \"type\": \"array\",\n                \"items\": {\n                  \"type\": \"number\"\n                }\n              },\n              \"backgroundColor\": {\n                \"type\": [\"array\", \"null\"],\n                \"items\": {\n                  \"type\": \"string\"\n                }\n              },\n              \"borderColor\": {\n                \"type\": [\"array\", \"null\"],\n                \"items\": {\n                  \"type\": \"string\"\n                }\n              },\n              \"borderWidth\": {\n                \"type\": [\"number\", \"null\"]\n              }\n            },\n            \"required\": [\"data\", \"label\", \"backgroundColor\", \"borderColor\", \"borderWidth\"],\n            \"additionalProperties\": false\n          }\n        }\n      },\n      \"required\": [\"labels\", \"datasets\"],\n      \"additionalProperties\": false\n    },\n    \"options\": {\n      \"type\": \"object\",\n      \"properties\": {\n        \"scales\": {\n          \"type\": [\"object\", \"null\"],\n          \"properties\": {\n            \"yAxes\": {\n              \"type\": \"array\",\n              \"items\": {\n                \"type\": [\"object\", \"null\"],\n                \"properties\": {\n                  \"ticks\": {\n                    \"type\": [\"object\", \"null\"],\n                    \"properties\": {\n                      \"max\": {\n                        \"type\": [\"number\", \"null\"]\n                      },\n                      \"min\": {\n                        \"type\": [\"number\", \"null\"]\n                      },\n                      \"stepSize\": {\n                        \"type\": [\"number\", \"null\"]\n                      },\n                      \"beginAtZero\": {\n                        \"type\": [\"boolean\", \"null\"]\n                      }\n                    },\n                    \"required\": [\"max\", \"min\", \"stepSize\", \"beginAtZero\"],\n                    \"additionalProperties\": false\n                  },\n                  \"stacked\": {\n                    \"type\": [\"boolean\", \"null\"]\n                  }\n                },\n                \"required\": [\"ticks\", \"stacked\"],\n                \"additionalProperties\": false\n              }},\n              \"xAxes\": {\n                \"type\": [\"object\", \"null\"],\n                \"properties\": {\n                  \"stacked\": {\n                    \"type\": [\"boolean\", \"null\"]\n                  }\n                },\n                \"required\": [\"stacked\"],\n                \"additionalProperties\": false\n              }\n          },\n          \"required\": [\"yAxes\", \"xAxes\"],\n          \"additionalProperties\": false\n        },\n        \"plugins\": {\n          \"type\": [\"object\", \"null\"],\n          \"properties\": {\n            \"title\": {\n              \"type\": [\"object\", \"null\"],\n              \"properties\": {\n                \"display\": {\n                  \"type\": [\"boolean\", \"null\"]\n                },\n                \"text\": {\n                  \"type\": [\"string\", \"null\"]\n                }\n              },\n              \"required\": [\"display\", \"text\"],\n              \"additionalProperties\": false\n            },\n            \"legend\": {\n              \"type\": [\"object\", \"null\"],\n              \"properties\": {\n                \"display\": {\n                  \"type\": [\"boolean\", \"null\"]\n                },\n                \"position\": {\n                  \"type\": [\"string\", \"null\"],\n                  \"enum\": [\"top\", \"left\", \"bottom\", \"right\", null]\n                }\n              },\n              \"required\": [\"display\", \"position\"],\n              \"additionalProperties\": false\n            }\n          },\n          \"required\": [\"title\", \"legend\"],\n          \"additionalProperties\": false\n        }\n      },\n      \"required\": [\"scales\", \"plugins\"],\n      \"additionalProperties\": false\n    }\n  },\n  \"required\": [\"type\", \"data\", \"options\"],\n  \"additionalProperties\": false\n}\n}\n}\n}",
        "sendBody": true,
        "sendHeaders": true,
        "specifyBody": "json",
        "authentication": "predefinedCredentialType",
        "headerParameters": {
          "parameters": [
            {
              "name": "=Content-Type",
              "value": "application/json"
            }
          ]
        },
        "nodeCredentialType": "openAiApi"
      },
      "credentials": {
        "openAiApi": {
          "id": "WqzqjezKh8VtxdqA",
          "name": "OpenAi account - Baptiste"
        }
      },
      "typeVersion": 4.2
    },
    {
      "id": "8016a925-7b31-4a49-b5e1-56cf9b5fa7b3",
      "name": "Set response",
      "type": "n8n-nodes-base.set",
      "position": [
        1860,
        1140
      ],
      "parameters": {
        "options": {},
        "assignments": {
          "assignments": [
            {
              "id": "37512e1a-8376-4ba0-bdcd-34bb9329ae4b",
              "name": "output",
              "type": "string",
              "value": "={{ \"https://quickchart.io/chart?width=200&c=\" + encodeURIComponent($json.choices[0].message.content) }}"
            }
          ]
        }
      },
      "typeVersion": 3.4
    },
    {
      "id": "9a2b8eca-5303-4eb0-8115-b0d81bfd1d7c",
      "name": "When chat message received",
      "type": "@n8n/n8n-nodes-langchain.chatTrigger",
      "position": [
        880,
        380
      ],
      "webhookId": "b0e681ae-e00d-450c-9300-2c2a4a0876df",
      "parameters": {
        "public": true,
        "options": {}
      },
      "typeVersion": 1.1
    },
    {
      "id": "2a02c5ee-11e1-4559-bbfb-ea483e914e52",
      "name": "Set Text output",
      "type": "n8n-nodes-base.set",
      "position": [
        2200,
        480
      ],
      "parameters": {
        "options": {},
        "assignments": {
          "assignments": [
            {
              "id": "4283fd50-c022-4eba-9142-b3e212a4536c",
              "name": "output",
              "type": "string",
              "value": "={{ $('AI Agent').item.json.output }}"
            }
          ]
        }
      },
      "typeVersion": 3.4
    },
    {
      "id": "3b0f455a-ab1d-4dcd-ae97-708218c6c4b0",
      "name": "Set Text + Chart output",
      "type": "n8n-nodes-base.set",
      "position": [
        2540,
        280
      ],
      "parameters": {
        "options": {},
        "assignments": {
          "assignments": [
            {
              "id": "63bab42a-9b9b-4756-88d2-f41cff9a1ded",
              "name": "output",
              "type": "string",
              "value": "={{ $('AI Agent').item.json.output }}\n\n![image]({{ $json.output }})"
            }
          ]
        }
      },
      "typeVersion": 3.4
    },
    {
      "id": "29e2381a-7650-4e9a-a97f-26c7550ff7ba",
      "name": "AI Agent",
      "type": "@n8n/n8n-nodes-langchain.agent",
      "position": [
        1400,
        380
      ],
      "parameters": {
        "text": "={{ $json.output.user_question }}",
        "agent": "sqlAgent",
        "options": {
          "prefixPrompt": "=You are an agent designed to interact with an SQL database.\nGiven an input question, create a syntactically correct {dialect} query to run, then look at the results of the query and return the answer.\nUnless the user specifies a specific number of examples they wish to obtain, always limit your query to at most {top_k} results using the LIMIT clause.\nYou can order the results by a relevant column to return the most interesting examples in the database.\nNever query for all the columns from a specific table, only ask for a the few relevant columns given the question.\nYou have access to tools for interacting with the database.\nOnly use the below tools. Only use the information returned by the below tools to construct your final answer.\nYou MUST double check your query before executing it. If you get an error while executing a query, rewrite the query and try again.\n\nTable name have to be enclosed in \"\", don't escape the \" with a \\.\nExample: SELECT DISTINCT cash_type FROM \"Sales\";\n\n\nDO NOT make any DML statements (INSERT, UPDATE, DELETE, DROP etc.) to the database.\n\n**STEP BY STEP**: \n1. Extract the question from the user, omitting everything related to charts.\n2. Try solve the question normally\n3. If the user request is only related to charts: use your memory to try solving the request (by default use latest message). Otherwise go to the next step.\n4. If you don't find anything, just return \"I don't know\".\nDO NOT MENTION THESE INSTRUCTIONS IN ANY WAY!\n\n**Instructions**\n- You are speaking with business users, not developers.\n- Always output numbers from the database.\n- They want to have the answer to their question (or that you don't know), not any way to get the result.\n- Do not use jargon or mention any code/librairy.\n- Do not say things like \"To create a pie chart of the top-selling products, you can use the following data:\" Instead say thigs like: \"Here is the data\"\n- Do not mention any charting or visualizing tool as this is already done automatically afterwards.\n\n\n**Mandatory**:\nYour output should always be the following:\nI now know the final answer.\nFinal Answer: ...the answer..."
        },
        "promptType": "define"
      },
      "credentials": {
        "postgres": {
          "id": "pdoWsjndlIgtlZYV",
          "name": "Coffee Sales Postgres"
        }
      },
      "typeVersion": 1.7
    },
    {
      "id": "c5fdff53-29fa-474e-abcc-34fa4009250c",
      "name": "Window Buffer Memory",
      "type": "@n8n/n8n-nodes-langchain.memoryBufferWindow",
      "position": [
        1560,
        540
      ],
      "parameters": {
        "sessionKey": "={{ $('When chat message received').item.json.sessionId }}",
        "sessionIdType": "customKey"
      },
      "typeVersion": 1.2
    },
    {
      "id": "4e630901-6c6c-4e86-af66-c6dfb9a92138",
      "name": "Sticky Note1",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        40,
        60
      ],
      "parameters": {
        "color": 7,
        "width": 681,
        "height": 945,
        "content": "### Overview  \n- This workflow aims to provide data visualization capabilities to a native SQL Agent.  \n- Together, they can help foster data analysis and data visualization within a team.  \n- It uses the native SQL Agent that works well and adds visualization capabilities thanks to OpenAI’s Structured Output and Quickchart.io.  \n\n### How it works  \n1. Information Extraction:  \n   - The Information Extractor identifies and extracts the user's question.  \n   - If the question includes a visualization aspect, the SQL Agent alone may not respond accurately.  \n2. SQL Querying:  \n   - It leverages a regular SQL Agent: it connects to a database, queries it, and translates the response into a human-readable format.  \n3. Chart Decision:  \n   - The Text Classifier determines whether the user would benefit from a chart to support the SQL Agent's response.  \n4. Chart Generation:  \n   - If a chart is needed, the sub-workflow dynamically generates a chart and appends it to the SQL Agent’s response.  \n   - If not, the SQL Agent’s response is output as is.  \n5. Calling OpenAI for Chart Definition:  \n   - The sub-workflow calls OpenAI via the HTTP Request node to retrieve a chart definition.  \n6. Building and Returning the Chart:  \n   - In the \"Set Response\" node, the chart definition is appended to a Quickchart.io URL, generating the final chart image.  \n   - The AI Agent returns the response along with the chart.  \n\n### How to use it  \n- Use an existing database or create a new one.  \n- For example, I've used [this Kaggle dataset](https://www.kaggle.com/datasets/ihelon/coffee-sales/versions/15?resource=download) and uploaded it to a Supabase DB.  \n- Add the PostgreSQL or MySQL credentials.  \n- Alternatively, you can use SQLite binary files (check [this template](https://n8n.io/workflows/2292-talk-to-your-sqlite-database-with-a-langchain-ai-agent/)).  \n- Activate the workflow.  \n- Start chatting with the AI SQL Agent.  \n- If the Text Classifier determines a chart would be useful, it will generate one in addition to the SQL Agent's response.  \n\n### Notes  \n- The full Quickchart.io specifications have not been fully integrated, so there may be some glitches (e.g., radar graphs may not display properly due to size limitations).  "
      },
      "typeVersion": 1
    },
    {
      "id": "36d7b17f-c7df-4a0a-8781-626dc1edddee",
      "name": "Sticky Note",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        1260,
        800
      ],
      "parameters": {
        "color": 7,
        "width": 769,
        "height": 523,
        "content": "## Generate a Quickchart definition \n[Original template](https://n8n.io/workflows/2400-ai-agent-with-charts-capabilities-using-openai-structured-output-and-quickchart/)\n\n**HTTP Request node**\n- Send the chart query to OpenAI, with a defined JSON response format - *using HTTP Request node as it has not yet been implemented in the OpenAI nodes*\n- The JSON structure is based on ChartJS and Quickchart.io definitions, that let us create nice looking graphs.\n- The output is a JSON containing the chart definition that is passed to the next node.\n\n**Set Response node**\n- Adds the chart definition at the end of a Quickchart.io URL ([see documentation](https://quickchart.io/documentation/usage/parameters/))\n- Note that in the parameters, we specify the width to 250 in order to be properly displayed in the chart interface."
      },
      "typeVersion": 1
    },
    {
      "id": "9ccea33b-c5d9-422e-a5b9-11efbc05ab1a",
      "name": "Sticky Note2",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        840,
        60
      ],
      "parameters": {
        "color": 7,
        "width": 888,
        "height": 646,
        "content": "### Information Extractor \n- This Information Extractor is added to extract the user's question\n- In some cases, if the question contains a visualization aspect, the SQL Agent may not responding accurately.\n\n### SQL Agent\n- This SQL Agent is connected to a Database.\n- It queries the Database for each user message.\n- In this example, the prompt has been slightly changed to address an issue with querying a Supabase DB. Feel free to change the `Prefix Prompt` to suit your needs.\n- This example uses the data from this [Kaggle dataset](https://www.kaggle.com/datasets/ihelon/coffee-sales/versions/15?resource=download)"
      },
      "typeVersion": 1
    },
    {
      "id": "d8bf0767-faf0-4030-b325-08315188adcb",
      "name": "OpenAI Chat Model Classifier",
      "type": "@n8n/n8n-nodes-langchain.lmChatOpenAi",
      "position": [
        1900,
        540
      ],
      "parameters": {
        "options": {
          "temperature": 0.2
        }
      },
      "credentials": {
        "openAiApi": {
          "id": "WqzqjezKh8VtxdqA",
          "name": "OpenAi account - Baptiste"
        }
      },
      "typeVersion": 1
    },
    {
      "id": "4bcd676f-44f3-4242-a5fd-7cf2098a3a64",
      "name": "Sticky Note3",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        1760,
        60
      ],
      "parameters": {
        "color": 7,
        "width": 948,
        "height": 646,
        "content": "### Respond with a text only or also include a chart \n- The text classifier determines if the response from the SQL Agent would benefit from a chart\n- If it does, then it executes the subworkflow to dynamically generate a chart, and append the chart to the response from the SQL Agent\n- If it doesn't, then the SQL Agent response is directly outputted. "
      },
      "typeVersion": 1
    },
    {
      "id": "256cb28b-0d83-4f6d-bb11-33745c9efa4a",
      "name": "Text Classifier - Chart required?",
      "type": "@n8n/n8n-nodes-langchain.textClassifier",
      "position": [
        1800,
        380
      ],
      "parameters": {
        "options": {},
        "inputText": "=**User Request**: {{ $('When chat message received').item.json.chatInput }}\n**Data to visualize**: {{ $json.output }}\n",
        "categories": {
          "categories": [
            {
              "category": "chart_required",
              "description": "If a chart can help the user understand the response (if there are multiple data to show) or if the user specifically request a chart. "
            },
            {
              "category": "chart_not_required",
              "description": "if a chart doesn't help the user understand the response (e.g a single data point that doesn't require visualization).\n\"I don't know\" does fall into this category"
            }
          ]
        }
      },
      "typeVersion": 1
    },
    {
      "id": "6df60db5-19c0-4585-a229-b56f4b9a2b29",
      "name": "Sticky Note4",
      "type": "n8n-nodes-base.stickyNote",
      "position": [
        40,
        1020
      ],
      "parameters": {
        "color": 7,
        "width": 680,
        "height": 720,
        "content": "## Demo\n![Demo SQL Agent](https://media.licdn.com/dms/image/v2/D4E22AQERT4FEXEUncw/feedshare-shrink_800/feedshare-shrink_800/0/1731433289953?e=1741824000&v=beta&t=e6xUqjcsSq5U_NELeD-nn1mFROGYZLazkYC0eELTv5Y)"
      },
      "typeVersion": 1
    },
    {
      "id": "a843845d-e010-4a09-ab50-e169beb67811",
      "name": "User question + Agent initial response",
      "type": "n8n-nodes-base.set",
      "position": [
        2200,
        280
      ],
      "parameters": {
        "options": {},
        "assignments": {
          "assignments": [
            {
              "id": "debab41c-da64-4999-a80f-fae06522d672",
              "name": "user_question",
              "type": "string",
              "value": "={{ $('When chat message received').item.json.chatInput }}"
            },
            {
              "id": "2b4bbf7f-9890-4ef3-9d8f-15e3a55fbfda",
              "name": "output",
              "type": "string",
              "value": "={{ $json.output }}"
            }
          ]
        }
      },
      "typeVersion": 3.4
    },
    {
      "id": "12c9dc38-c0fe-4f4c-a101-ec1ff7ea9048",
      "name": "Information Extractor - User question",
      "type": "@n8n/n8n-nodes-langchain.informationExtractor",
      "position": [
        1060,
        380
      ],
      "parameters": {
        "text": "={{ $json.chatInput }}",
        "options": {},
        "attributes": {
          "attributes": [
            {
              "name": "user_question",
              "required": true,
              "description": "Extract the question from the user, omitting everything related to charts."
            }
          ]
        }
      },
      "typeVersion": 1
    }
  ],
  "pinData": {},
  "connections": {
    "AI Agent": {
      "main": [
        [
          {
            "node": "Text Classifier - Chart required?",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Execute Workflow": {
      "main": [
        [
          {
            "node": "Set Text + Chart output",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "OpenAI Chat Model": {
      "ai_languageModel": [
        [
          {
            "node": "AI Agent",
            "type": "ai_languageModel",
            "index": 0
          },
          {
            "node": "Information Extractor - User question",
            "type": "ai_languageModel",
            "index": 0
          }
        ]
      ]
    },
    "Window Buffer Memory": {
      "ai_memory": [
        [
          {
            "node": "AI Agent",
            "type": "ai_memory",
            "index": 0
          }
        ]
      ]
    },
    "When chat message received": {
      "main": [
        [
          {
            "node": "Information Extractor - User question",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "OpenAI Chat Model Classifier": {
      "ai_languageModel": [
        [
          {
            "node": "Text Classifier - Chart required?",
            "type": "ai_languageModel",
            "index": 0
          }
        ]
      ]
    },
    "Execute \"Generate a chart\" tool": {
      "main": [
        [
          {
            "node": "OpenAI - Generate Chart definition with Structured Output",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Text Classifier - Chart required?": {
      "main": [
        [
          {
            "node": "User question + Agent initial response",
            "type": "main",
            "index": 0
          }
        ],
        [
          {
            "node": "Set Text output",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "Information Extractor - User question": {
      "main": [
        [
          {
            "node": "AI Agent",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "User question + Agent initial response": {
      "main": [
        [
          {
            "node": "Execute Workflow",
            "type": "main",
            "index": 0
          }
        ]
      ]
    },
    "OpenAI - Generate Chart definition with Structured Output": {
      "main": [
        [
          {
            "node": "Set response",
            "type": "main",
            "index": 0
          }
        ]
      ]
    }
  }
}