From c50f0f31f418037063bd97efcdc0f60b89594212 Mon Sep 17 00:00:00 2001 From: Lain Soykaf Date: Tue, 14 May 2024 16:56:58 +0400 Subject: Docs/Search: Add basic documentation of the qdrant search --- docs/configuration/search.md | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'docs/configuration') diff --git a/docs/configuration/search.md b/docs/configuration/search.md index 0316c9bf4..682d1e52a 100644 --- a/docs/configuration/search.md +++ b/docs/configuration/search.md @@ -10,6 +10,12 @@ To use built-in search that has no external dependencies, set the search module While it has no external dependencies, it has problems with performance and relevancy. +## QdrantSearch + +This uses the vector search engine [Qdrant](https://qdrant.tech) to search the posts in a vector space. This needs a way to generate embeddings, for now only the [Ollama](Ollama) api is supported. + +The default settings will support a setup where both Ollama and Qdrant run on the same system as pleroma. The embedding model used by Ollama will need to be pulled first (e.g. `ollama pull snowflake-arctic-embed:xs`) for the embedding to work. + ## Meilisearch Note that it's quite a bit more memory hungry than PostgreSQL (around 4-5G for ~1.2 million -- cgit v1.2.3 From e142ea400a9ed3595f8d432edd90ea26fc7d2eb5 Mon Sep 17 00:00:00 2001 From: Lain Soykaf Date: Sun, 19 May 2024 12:42:08 +0400 Subject: Docs: Switch docs from Ollama to OpenAI. --- docs/configuration/search.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'docs/configuration') diff --git a/docs/configuration/search.md b/docs/configuration/search.md index 682d1e52a..388f5acd1 100644 --- a/docs/configuration/search.md +++ b/docs/configuration/search.md @@ -12,9 +12,9 @@ While it has no external dependencies, it has problems with performance and rele ## QdrantSearch -This uses the vector search engine [Qdrant](https://qdrant.tech) to search the posts in a vector space. This needs a way to generate embeddings, for now only the [Ollama](Ollama) api is supported. +This uses the vector search engine [Qdrant](https://qdrant.tech) to search the posts in a vector space. This needs a way to generate embeddings and uses the [OpenAI API](https://platform.openai.com/docs/guides/embeddings/what-are-embeddings). This is implemented by several project besides OpenAI itself, including the python-based fastembed-server found in `supplemental/search/fastembed-api`. -The default settings will support a setup where both Ollama and Qdrant run on the same system as pleroma. The embedding model used by Ollama will need to be pulled first (e.g. `ollama pull snowflake-arctic-embed:xs`) for the embedding to work. +The default settings will support a setup where both the fastembed server and Qdrant run on the same system as pleroma. ## Meilisearch -- cgit v1.2.3 From 6a3a0cc0f5995185428c92f3c53e9c8524ea6856 Mon Sep 17 00:00:00 2001 From: Lain Soykaf Date: Sun, 19 May 2024 13:20:37 +0400 Subject: Docs: Write docs for the QdrantSearch --- docs/configuration/search.md | 20 +++++++++++++++++++- 1 file changed, 19 insertions(+), 1 deletion(-) (limited to 'docs/configuration') diff --git a/docs/configuration/search.md b/docs/configuration/search.md index 388f5acd1..6598e533f 100644 --- a/docs/configuration/search.md +++ b/docs/configuration/search.md @@ -14,7 +14,25 @@ While it has no external dependencies, it has problems with performance and rele This uses the vector search engine [Qdrant](https://qdrant.tech) to search the posts in a vector space. This needs a way to generate embeddings and uses the [OpenAI API](https://platform.openai.com/docs/guides/embeddings/what-are-embeddings). This is implemented by several project besides OpenAI itself, including the python-based fastembed-server found in `supplemental/search/fastembed-api`. -The default settings will support a setup where both the fastembed server and Qdrant run on the same system as pleroma. +The default settings will support a setup where both the fastembed server and Qdrant run on the same system as pleroma. To use it, set the search provider and run the fastembed server, see the README in `supplemental/search/fastembed-api`: + +https://qdrant.github.io/fastembed/examples/Supported_Models/ + +> config :pleroma, Pleroma.Search, module: Pleroma.Search.QdrantSearch + +You will also need to create the Qdrant index once by running `mix pleroma.search.indexer create_index`. Running `mix pleroma.search.indexer index` will retroactively index the last 100_000 activities. + +### Indexing and model options + +To see the available configuration options, check out the QdrantSearch section in `config/config.exs`. + +The default indexing option work for the default model (`snowflake-arctic-embed-xs`). To optimize for a low memory footprint, adjust the index configuration as described in the [Qdrant docs](https://qdrant.tech/documentation/guides/optimize/). + +Different embedding models will need different vector size settings. You can see a list of the models supported by the fastembed server [here](https://qdrant.github.io/fastembed/examples/Supported_Models), including their vector dimensions. These vector dimensions need to be set in the `qdrant_index_configuration`. + +E.g, If you want to use `sentence-transformers/all-MiniLM-L6-v2` as a model, you will not need to adjust things, because it and `snowflake-arctic-embed-xs` are both 384 dimensional models. If you want to use `snowflake/snowflake-arctic-embed-l`, you will need to adjust the `size` parameter in the `qdrant_index_configuration` to 1024, as it has a dimension of 1024. + +When using a different model, you will need do drop the index and recreate it (`mix pleroma.search.indexer drop_index` and `mix pleroma.search.indexer create_index`), as the different embeddings are not compatible with each other. ## Meilisearch -- cgit v1.2.3 From 6ec306d0684f3c5c05d768a3c431008925f21f15 Mon Sep 17 00:00:00 2001 From: Lain Soykaf Date: Sun, 19 May 2024 13:24:24 +0400 Subject: Docs: Add more information about index memory consumption. --- docs/configuration/search.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'docs/configuration') diff --git a/docs/configuration/search.md b/docs/configuration/search.md index 6598e533f..ed85acd2a 100644 --- a/docs/configuration/search.md +++ b/docs/configuration/search.md @@ -26,7 +26,7 @@ You will also need to create the Qdrant index once by running `mix pleroma.searc To see the available configuration options, check out the QdrantSearch section in `config/config.exs`. -The default indexing option work for the default model (`snowflake-arctic-embed-xs`). To optimize for a low memory footprint, adjust the index configuration as described in the [Qdrant docs](https://qdrant.tech/documentation/guides/optimize/). +The default indexing option work for the default model (`snowflake-arctic-embed-xs`). To optimize for a low memory footprint, adjust the index configuration as described in the [Qdrant docs](https://qdrant.tech/documentation/guides/optimize/). See also [this blog post](https://qdrant.tech/articles/memory-consumption/) that goes into detail. Different embedding models will need different vector size settings. You can see a list of the models supported by the fastembed server [here](https://qdrant.github.io/fastembed/examples/Supported_Models), including their vector dimensions. These vector dimensions need to be set in the `qdrant_index_configuration`. -- cgit v1.2.3 From dbaab6f54e306e5fb930ce1ed0699631c8aeaae1 Mon Sep 17 00:00:00 2001 From: Lain Soykaf Date: Sun, 19 May 2024 13:38:31 +0400 Subject: Docs: Mention running the Qdrant server --- docs/configuration/search.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'docs/configuration') diff --git a/docs/configuration/search.md b/docs/configuration/search.md index ed85acd2a..d34f84d4f 100644 --- a/docs/configuration/search.md +++ b/docs/configuration/search.md @@ -16,10 +16,10 @@ This uses the vector search engine [Qdrant](https://qdrant.tech) to search the p The default settings will support a setup where both the fastembed server and Qdrant run on the same system as pleroma. To use it, set the search provider and run the fastembed server, see the README in `supplemental/search/fastembed-api`: -https://qdrant.github.io/fastembed/examples/Supported_Models/ - > config :pleroma, Pleroma.Search, module: Pleroma.Search.QdrantSearch +Then, start the Qdrant server, see [here](https://qdrant.tech/documentation/quick-start/) for instructions. + You will also need to create the Qdrant index once by running `mix pleroma.search.indexer create_index`. Running `mix pleroma.search.indexer index` will retroactively index the last 100_000 activities. ### Indexing and model options -- cgit v1.2.3 From fc7ce339edc40cb791d321a20f01f2568337b845 Mon Sep 17 00:00:00 2001 From: Lain Soykaf Date: Mon, 27 May 2024 21:28:20 +0400 Subject: Cheatsheet: Add allow_unsafe_eval --- docs/configuration/cheatsheet.md | 1 + 1 file changed, 1 insertion(+) (limited to 'docs/configuration') diff --git a/docs/configuration/cheatsheet.md b/docs/configuration/cheatsheet.md index ca2ce6369..78997c4db 100644 --- a/docs/configuration/cheatsheet.md +++ b/docs/configuration/cheatsheet.md @@ -472,6 +472,7 @@ This will make Pleroma listen on `127.0.0.1` port `8080` and generate urls start * ``ct_max_age``: The maximum age for the `Expect-CT` header if sent. * ``referrer_policy``: The referrer policy to use, either `"same-origin"` or `"no-referrer"`. * ``report_uri``: Adds the specified url to `report-uri` and `report-to` group in CSP header. +* `allow_unsafe_eval`: Adds `wasm-unsafe-eval` to the CSP header. Needed for some non-essential frontend features like Flash emulation. ### Pleroma.Web.Plugs.RemoteIp -- cgit v1.2.3