[Question]: Error after fetching knowledge graph using https api

[cyx96]

Self Checks

• I have searched for existing issues search for existing issues, including closed ones.
• I confirm that I am using English to submit this report (Language Policy).
• Non-english title submitions will be closed directly ( 非英文标题的提交将会被直接关闭 ) (Language Policy).
• Please do not modify this template :) and fill in all the required fields.

Describe your problem

I’ve enabled the knowledge graph on my dataset, as confirmed by the list-datasets endpoint:

{'code': 0,
 'data': [{'avatar': '',
   'chunk_count': 724,
   'chunk_method': 'naive',
   'create_date': 'Thu, 24 Jul 2025 13:09:23 GMT',
   'create_time': 1753333763643,
   'created_by': 'c56b72d2680211f081db024cd4240d23',
   'description': None,
   'document_count': 1,
   'embedding_model': 'mxbai-embed-large:335m@Ollama',
   'id': '5d0d1e44684c11f084f49a9a55bc04ae',
   'language': 'English',
   'name': 'History',
   'pagerank': 0,
   'parser_config': {'auto_keywords': 0,
    'auto_questions': 0,
    'chunk_token_num': 128,
    'delimiter': '\n',
    'graphrag': {'community': True,
     'entity_types': ['organization', 'person', 'geo', 'event', 'category'],
     'method': 'light',
     'resolution': True,
     'use_graphrag': True},
    'html4excel': False,
    'layout_recognize': 'DeepDOC',
    'raptor': {'use_raptor': False}},
   'permission': 'me',
   'similarity_threshold': 0.2,
   'status': '1',
   'tenant_id': 'c56b72d2680211f081db024cd4240d23',
   'token_num': 110208,
   'update_date': 'Wed, 30 Jul 2025 02:17:50 GMT',
   'update_time': 1753813070212,
   'vector_similarity_weight': 0.3}]}

However, when I try to fetch the graph with:

curl -X GET \
  "http://{address}/api/v1/datasets/5d0d1e44684c11f084f49a9a55bc04ae/knowledge_graph" \
  -H "Authorization: Bearer <YOUR_API_KEY>"

I get a 404:

2025-07-30 07:02:54,038 ERROR    20 404 Not Found: The requested URL was not found on the server. If you entered the URL manually please check your spelling and try again.
Traceback (most recent call last):
  File "/ragflow/.venv/lib/python3.10/site-packages/flask/app.py", line 880, in full_dispatch_request
    rv = self.dispatch_request()
  File "/ragflow/.venv/lib/python3.10/site-packages/flask/app.py", line 854, in dispatch_request
    self.raise_routing_exception(req)
  File "/ragflow/.venv/lib/python3.10/site-packages/flask/app.py", line 463, in raise_routing_exception
    raise request.routing_exception  # type: ignore[misc]
  File "/ragflow/.venv/lib/python3.10/site-packages/flask/ctx.py", line 362, in match_request
    result = self.url_adapter.match(return_rule=True)  # type: ignore
  File "/ragflow/.venv/lib/python3.10/site-packages/werkzeug/routing/map.py", line 629, in match
    raise NotFound() from None
werkzeug.exceptions.NotFound: 404 Not Found: The requested URL was not found on the server. If you entered the URL manually please check your spelling and try again.

I’ve double-checked the endpoint and spelling, which should be correct. Any idea what might be going wrong or whether I’m missing a configuration step? Thanks in advance for any advice!

[dosubot]

Hi @cyx96! I'm Dosu and I’m helping the ragflow team.

The 404 error happens because the /api/v1/datasets/<dataset_id>/knowledge_graph endpoint is only available in RAGFlow versions that include the knowledge graph HTTP API, which was added in mid-to-late July 2025. If your deployment predates these changes, the endpoint simply doesn't exist yet, even if the feature is enabled in your dataset config [source].

To resolve this, check your RAGFlow version. If it's older than the July 2025 releases, upgrade to a version after July 18, 2025. You can upgrade by editing your .env file to set the new version (for example, 0.18.0 or later), then running docker compose -f docker-compose.yml up -d—this will preserve your existing data [source]. After upgrading, delete any existing knowledge graph for your dataset and re-parse or re-upload your document to regenerate the graph, as the knowledge graph is not persistent across upgrades or restarts [source].

No extra environment variables or config flags are needed—just make sure you're running a recent enough version. If you still get a 404 after upgrading, double-check the logs and confirm the endpoint is present in your deployment. Let me know if you need help with any of these steps!

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}    

[cyx96]

@dosu Thanks for your swift and helpful response. You are a good bot!

However, I'm using RAGFLOW_IMAGE=infiniflow/ragflow:v0.19.1, which should be the latest version according to the RagFlow doc.

[dosubot]

Thanks for clarifying your version! There’s a known issue where the .env file in some v0.19.1 source packages still points to the old Docker image (v0.19.0-slim), so even if you set RAGFLOW_IMAGE=infiniflow/ragflow:v0.19.1, your server might actually be running the older image. This would explain the missing endpoint and 404 error [source].

To check, run docker ps and look at the IMAGE column for your running container—it should show infiniflow/ragflow:v0.19.1 (or v0.19.1-slim). If not, update the RAGFLOW_IMAGE line in docker/.env to the correct version, then run:

docker compose -f docker/docker-compose.yml pull
docker compose -f docker/docker-compose.yml up -d

This will ensure you’re running the right image [source].

Also, in v0.19.1, the knowledge graph isn’t persistent across restarts—after any restart or image update, you’ll need to regenerate the graph by re-uploading and parsing a file in your dataset for the endpoint to return data [source].

No extra config or flags are needed for the endpoint in v0.19.1. If you’re still seeing a 404 after these steps, let me know what you see in the logs or if the container image matches—happy to help debug further!

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}    

[cyx96]

@dosu Alright, I did what you asked and the following are my new updates:

1. Running docker ps I get:

CONTAINER ID   IMAGE                                              COMMAND                  CREATED      STATUS                 PORTS                                                                                                                       NAMES
4228397d14c1   infiniflow/ragflow:v0.19.1                         "./entrypoint.sh"        5 days ago   Up 6 hours             0.0.0.0:80->80/tcp, [::]:80->80/tcp, 0.0.0.0:443->443/tcp, [::]:443->443/tcp, 0.0.0.0:9380->9380/tcp, [::]:9380->9380/tcp   ragflow-server
3c754ef31bfe   mysql:8.0.39                                       "docker-entrypoint.s…"   5 days ago   Up 6 hours (healthy)   33060/tcp, 0.0.0.0:5455->3306/tcp, [::]:5455->3306/tcp                                                                      ragflow-mysql
4957f8327f87   elasticsearch:8.11.3                               "/bin/tini -- /usr/l…"   6 days ago   Up 6 hours (healthy)   9300/tcp, 0.0.0.0:1200->9200/tcp, [::]:1200->9200/tcp                                                                       ragflow-es-01
8ea0b4d92584   quay.io/minio/minio:RELEASE.2023-12-20T01-00-02Z   "/usr/bin/docker-ent…"   6 days ago   Up 6 hours             0.0.0.0:9000-9001->9000-9001/tcp, [::]:9000-9001->9000-9001/tcp                                                             ragflow-minio
a901f4a58a5e   valkey/valkey:8                                    "docker-entrypoint.s…"   6 days ago   Up 6 hours             0.0.0.0:6379->6379/tcp, [::]:6379->6379/tcp                                                                                 ragflow-redis

where you can clearly see the version is the latest infiniflow/ragflow:v0.19.1.
2. I didn't restart nor any image update after creating this knowledge graph. So this should be irrelevant to my issue.

[dosubot]

You're definitely running the correct image, and the endpoint should be available by default in v0.19.1—there are no feature flags or extra config needed, and a missing knowledge graph would return empty data (HTTP 200), not a 404. A 404 here usually means the route isn't registered or something is interfering with routing.

A few things to check next:

• Look at your RAGFlow server logs right after startup and when you hit the endpoint—see if there are any errors or warnings about route registration or Flask/FastAPI routing.
• Try enabling API docs by setting ENABLE_DOCS=1 in your environment (e.g., in your .env or docker-compose.yml), then visit /docs or /redoc on your server. See if /api/v1/datasets/{dataset_id}/knowledge_graph is listed.
• Test a simple GET to /api/v1/datasets/{dataset_id} (replace with your actual dataset ID) to confirm other dataset endpoints work as expected.

If the endpoint is missing from the docs or you see errors in the logs, let me know what you find—this will help pinpoint if it's a deployment, routing, or version-specific bug.

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}    

[cyx96]

@dosu So the following is the logs right after startup:

        ____   ___    ______ ______ __               
       / __ \ /   |  / ____// ____// /____  _      __
      / /_/ // /| | / / __ / /_   / // __ \| | /| / /
     / _, _// ___ |/ /_/ // __/  / // /_/ /| |/ |/ / 
    /_/ |_|/_/  |_|\____//_/    /_/ \____/ |__/|__/                             

    
2025-07-30 07:49:15,509 INFO     20 RAGFlow version: v0.19.1 full
2025-07-30 07:49:15,509 INFO     20 project base: /ragflow
2025-07-30 07:49:15,509 INFO     20 Current configs, from /ragflow/conf/service_conf.yaml:
        ragflow: {'host': '0.0.0.0', 'http_port': 9380}
        mysql: {'name': 'rag_flow', 'user': 'root', 'password': '********', 'host': 'mysql', 'port': 3306, 'max_connections': 900, 'stale_timeout': 300}
        minio: {'user': 'rag_flow', 'password': '********', 'host': 'minio:9000'}
        es: {'hosts': 'http://es01:9200', 'username': 'elastic', 'password': '********'}
        os: {'hosts': 'http://opensearch01:9201', 'username': 'admin', 'password': '********'}
        infinity: {'uri': 'infinity:23817', 'db_name': 'default_db'}
        redis: {'db': 1, 'password': '********', 'host': 'redis:6379'}
2025-07-30 07:49:15,509 WARNING  20 SECURITY WARNING: Using auto-generated SECRET_KEY. Generated key: e45ab26cef45652db7c2eddc8d861e3e066f14fca0c3a558cde2dae943f7bcba
2025-07-30 07:49:15,509 INFO     20 Use Elasticsearch http://es01:9200 as the doc engine.
2025-07-30 07:49:15,600 INFO     20 GET http://es01:9200/ [status:200 duration:0.089s]
2025-07-30 07:49:15,605 INFO     20 HEAD http://es01:9200/ [status:200 duration:0.005s]
2025-07-30 07:49:15,605 INFO     20 Elasticsearch http://es01:9200 is healthy.
2025-07-30 07:49:15,609 WARNING  20 Load term.freq FAIL!
2025-07-30 07:49:15,611 WARNING  20 Realtime synonym is disabled, since no redis connection.
2025-07-30 07:49:15,614 WARNING  20 Load term.freq FAIL!
2025-07-30 07:49:15,617 WARNING  20 Realtime synonym is disabled, since no redis connection.
2025-07-30 07:49:15,617 INFO     20 MAX_CONTENT_LENGTH: 134217728
2025-07-30 07:49:15,617 INFO     20 MAX_FILE_COUNT_PER_USER: 0
2025-07-30 07:49:16,929 INFO     20 init web data success:1.255831003189087
2025-07-30 07:49:16,930 INFO     20 Recursively importing plugins from path `/ragflow/plugin/embedded_plugins`
2025-07-30 07:49:16,931 INFO     20 Loaded llm_tools plugin BadCalculatorPlugin version 1.0.0
2025-07-30 07:49:16,931 INFO     20 RAGFlow HTTP server start...
2025-07-30 07:49:16,933 INFO     20 WARNING: This is a development server. Do not use it in a production deployment. Use a production WSGI server instead.
 * Running on all addresses (0.0.0.0)
 * Running on http://127.0.0.1:9380
 * Running on http://172.18.0.3:9380
2025-07-30 07:49:16,933 INFO     20 Press CTRL+C to quit
2025-07-30 07:49:17,932 INFO     20 Starting update_progress thread (delayed)
2025-07-30 07:49:17,932 INFO     20 update_progress lock_value: 9cad86cc-8b19-4b04-a888-72c675e9dba0
2025-07-30 07:49:19,769 INFO     30 task_executor_4228397d14c1_0 log path: /ragflow/logs/task_executor_4228397d14c1_0.log, log levels: {'peewee': 'WARNING', 'pdfminer': 'WARNING', 'root': 'INFO'}
2025-07-30 07:49:19,770 INFO     30 
  ______           __      ______                     __
 /_  __/___ ______/ /__   / ____/  _____  _______  __/ /_____  _____
  / / / __ `/ ___/ //_/  / __/ | |/_/ _ \/ ___/ / / / __/ __ \/ ___/
 / / / /_/ (__  ) ,<    / /____>  </  __/ /__/ /_/ / /_/ /_/ / /
/_/  \__,_/____/_/|_|  /_____/_/|_|\___/\___/\__,_/\__/\____/_/
    
2025-07-30 07:49:19,770 INFO     30 TaskExecutor: RAGFlow version: v0.19.1 full
2025-07-30 07:49:19,770 WARNING  30 SECURITY WARNING: Using auto-generated SECRET_KEY. Generated key: ad477446623911459ca808148c97ce0b04dd994a779e0f990c8b9ad8362918b4
2025-07-30 07:49:19,771 INFO     30 Use Elasticsearch http://es01:9200 as the doc engine.
2025-07-30 07:49:19,778 INFO     30 GET http://es01:9200/ [status:200 duration:0.006s]
2025-07-30 07:49:19,780 INFO     30 HEAD http://es01:9200/ [status:200 duration:0.002s]
2025-07-30 07:49:19,781 INFO     30 Elasticsearch http://es01:9200 is healthy.
2025-07-30 07:49:19,784 WARNING  30 Load term.freq FAIL!
2025-07-30 07:49:19,787 WARNING  30 Realtime synonym is disabled, since no redis connection.
2025-07-30 07:49:19,791 WARNING  30 Load term.freq FAIL!
2025-07-30 07:49:19,793 WARNING  30 Realtime synonym is disabled, since no redis connection.
2025-07-30 07:49:19,793 INFO     30 MAX_CONTENT_LENGTH: 134217728
2025-07-30 07:49:19,793 INFO     30 MAX_FILE_COUNT_PER_USER: 0

Is there any thing related to my issue?

[dosubot]

Nothing in your startup logs indicates a problem with route registration or the knowledge graph API itself—everything looks normal, and the warnings about Redis and term.freq are unrelated. Since you're running the correct image, the most likely causes for a 404 are deployment or routing issues.

To diagnose further: set ENABLE_DOCS=1 in your .env or docker-compose.yml, restart the container, and check /docs or /redoc in your browser. If /api/v1/datasets/{dataset_id}/knowledge_graph is missing from the docs, the route isn't registered; if it's present, try hitting other dataset endpoints like /api/v1/datasets/{dataset_id} to confirm general API routing works. If the endpoint is present in docs but still returns 404, double-check your URL, port, and any reverse proxy or frontend config that might be interfering with routing. Instructions here and common deployment pitfalls here.

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}    

[FlowerWater1019]

Have you fix this problem?

[cyx96]

Nope but I think

curl -X GET \
  "http://{address}/api/v1/datasets/5d0d1e44684c11f084f49a9a55bc04ae/knowledge_graph" \
  -H "Authorization: Bearer <YOUR_API_KEY>"

is probably for the future versions of RAGFlow (even though devs have previewed it in the doc already). I'm trying to say that we are too early for this feature to roll out.

[FlowerWater1019]

Oh my god that is unfortunate. Emm do you know the directory of the knowledge graph in docker containers? @dosu

[dosubot]

There isn't a directory inside the RAGFlow Docker container where the knowledge graph is stored. Instead, knowledge graph data is saved in the backend document store—either InfinityDB or Elasticsearch, depending on your setup—not as files or folders in the container filesystem. If you want to inspect or back up the knowledge graph, you'll need to do that at the database level, not by looking for files inside the container itself. More details here: source.

To reply, just mention @dosu.

---

^{How did I do? Good | Irrelevant | Incorrect | Verbose | Hallucination | Report 🐛 | Other}    

[FlowerWater1019]

my setting in .env file is DOC_ENGINE=${DOC_ENGINE:-elasticsearch}. So can you give me an certain example about how to get a certain knowledge graph file through API or some other way? Suppose that you have finished a knowledge graph and have the dataset_id. @dosu

[FlowerWater1019]

@dosu could you give me such an example?