[Question]: Which slicing method in ragflow can be more semantic for slicing?

[Monkeycrop]

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 found that using the general slicing method often results in a complete sentence being split into two different segments, which is unreasonable. I would like to know which slicing method can ensure the integrity of the paragraph sentence, or whether it can be segmented based on semantic information. @dosu

[dosubot]

The most semantic slicing methods in RAGFlow are the SentenceSplitter and SemanticSplitter, which are designed to preserve sentence or paragraph integrity and leverage semantic information for chunking. These splitters are implemented in the rag/flow/splitter directory and are preferable to the general or length-based methods, which often split sentences and reduce coherence details.

To use a semantic or sentence-preserving slicing method, select SentenceSplitter or SemanticSplitter when configuring your dataset or document chunking options. This can typically be set at the dataset or document configuration level before parsing, and you may need to update the document config after upload if needed details.

If you want strict control over segmentation (e.g., by delimiter), you can adjust the chunk token size and delimiter settings, but these may still be overridden by the slicing method logic details.

For best semantic preservation, use SentenceSplitter or SemanticSplitter, and configure your chunking method at the dataset or document level.

To reply, just mention @dosu.

---

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

[Monkeycrop]

Sorry, I just discovered that Ragflow released a new version, so I updated the service. How do I restart Ragflow to make the new code take effect? @dosu

[dosubot]

To restart Ragflow and apply the new version, use Docker Compose commands if you're running Ragflow in Docker. Run:

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

This stops all containers and then starts them again, ensuring the updated code is loaded. If you updated the Docker image, make sure the RAGFLOW_IMAGE variable in docker/.env points to the new version before restarting see docs.

For local development (not Docker), stop any running backend processes (like ragflow_server.py), then start them again. For frontend/UI changes, rebuild assets and restart the frontend server.

Let me know if you use a different deployment method and need tailored instructions.

To reply, just mention @dosu.

---

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

[Monkeycrop]

I deployed using Docker, and I ran docker compose down and docker compose up -d directly in the Docker folder. However, there's no change in the logs. How can I verify that the update was successful? @dosu

[dosubot]

You can verify that the Ragflow update was successful in a few ways:

1. Check the container logs: After restarting, Ragflow logs its version at startup. Run:

   docker logs -f ragflow-server
   

   Look for a line like RAGFlow version: vX.Y.Z near the top of the logs to confirm the running version cite.

2. Use the API: Ragflow exposes a /version endpoint that returns the current version. After logging in, send a GET request to /version (e.g., via curl or Postman) to see the version string cite.

3. Check the Docker image: Run docker ps to see which image and tag your container is using. Make sure it matches the version you expect.

If you don’t see the new version, double-check that your docker/.env file points to the correct image tag, and try pulling the latest image with:

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

If issues persist, try stopping and removing volumes with docker compose -f docker/docker-compose.yml down -v before bringing the containers back up.

To reply, just mention @dosu.

---

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

[Monkeycrop]

Is the latest version of ragflow v0.20.3 slim?

I also have a question: if I re-pull the image, will I lose all my file data and user data? @dosu

[dosubot]

The latest version of Ragflow is not v0.20.3; v0.21.1 is currently documented as the latest, and both full and slim variants are available for recent versions (e.g., infiniflow/ragflow:v0.21.1-slim). The slim variant for v0.20.3 should exist as infiniflow/ragflow:v0.20.3-slim, following the same naming convention see docs.

Re-pulling the Docker image or updating to a new version will not erase your file data or user data as long as you do not use the -v flag with docker compose down. Ragflow stores persistent data in Docker volumes and external services (like MinIO, MySQL, Redis), which are not affected by image updates or container restarts. Data loss only occurs if you explicitly remove volumes (e.g., with docker compose down -v) details details.

To reply, just mention @dosu.

---

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

[Monkeycrop]

How should I update Ragflow to the latest version? I don't want to lose my existing data, including data from MySQL, Redis, Elasticsearch, and MiniIO, etc. @dosu

[dosubot]

To update Ragflow to the latest version without losing your data (including MySQL, Redis, Elasticsearch, MinIO), follow these steps:

1. Back up your important config files: /ragflow/conf/service_conf.yaml, docker/.env, and docker/service_conf.yaml.template reference.
2. Update the RAGFLOW_IMAGE variable in docker/.env to the latest version tag you want.
3. Run:

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

   This will upgrade Ragflow and preserve all your data reference.
4. Do not use docker compose down -v during the upgrade, as this will delete your data volumes and cause data loss reference.
5. For extra safety, you can back up your databases and files:
   • MySQL:

     docker exec ragflow-mysql mysqldump -uroot -pinfini_rag_flow rag_flow > ./ragflow-mysql.sql

   • Elasticsearch: Use elasticdump to export analyzers, mappings, and data for each index starting with ragflow_.
   • MinIO: Copy the volume data directory or use the MinIO client to mirror data to a backup directory reference.

Your persistent data will be preserved as long as you avoid removing Docker volumes. Double-check your configs after upgrading to ensure everything works as expected.

To reply, just mention @dosu.

---

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

[Monkeycrop]

docker compose -f docker/docker-compose.yml pull docker compose -f docker/docker-compose.yml up -d
I ran the code above, but it seems the ragflow version is still v0.20.3 slim.@dosu

[Magicbook1108]

1. Delete the original images
2. docker compose -f docker-compose-base.yml up -d
3. in docker/.env add SVR_WEB_HTTP_PORT=80
4. visit http://localhost:80