How to use postgrest with hivemind

 If you haven't been following the recent development with hivemind. We are actively moving away from the python-based webserver to an sql only implementation where the webserver is postgresql itself. This leads to massive performance improvements. In order to do this we use an awfully named tool: Postgres**t** (note the T at the end) This means that if you are a developer on hivemind and want to test an endpoint, you will need to do things a tad differently. ## Step 1: install postgrest Head on the official repo and download the version that corresponds to your needs: https://github.com/PostgREST/postgrest/releases/tag/v12.0.2 Right now we use 12.0.2, but that number may change. You can always check this [dockerfile](https://gitlab.syncad.com/hive/hivemind/-/blob/develop/Dockerfile?ref_type=heads#L5) to see if it's any different. Once you get it it's a simple binary, so just shove that in /usr/bin and you're good to go: ``` sudo tar -xf postgrest-v12.0.2-linux-static-x64.tar.xz -C /usr/bin # check if it was correctly installed postgrest --version ``` ## Step 2: launching the server Hivemind provides a handy script for that: https://gitlab.syncad.com/hive/hivemind/-/blob/develop/scripts/start_postgrest.sh?ref_type=heads just execute: ``` /home/howo/hivemind/scripts/start_postgrest.sh --postgres-url=postgresql://postgres:@localhost:5432/haf_block_log 03/Jun/2025:14:37:42 +0900: Starting PostgREST 12.0.2... 03/Jun/2025:14:37:42 +0900: Attempting to connect to the database... 03/Jun/2025:14:37:42 +0900: Connection successful 03/Jun/2025:14:37:42 +0900: Admin server listening on port 3001 03/Jun/2025:14:37:42 +0900: Listening on port 8080 03/Jun/2025:14:37:42 +0900: Config reloaded 03/Jun/2025:14:37:42 +0900: Listening for notifications on the pgrst channel 03/Jun/2025:14:37:42 +0900: Schema cache loaded ``` # Step 3: querying hivemind At this point you're good to go ! But you'll notice that if you try to call an endpoint: ``` curl 'http://localhost:8080' --compressed -X POST -H 'Content-Type: application/json' --data-raw '{"id":6,"jsonrpc":"2.0","method":"bridge.account_notifications","params":{"account":"howo","limit":1}}' | jq % Total % Received % Xferd Average Speed Time Time Time Current Dload Upload Total Spent Left Speed 100 110 0 2 100 108 1831 98901 --:--:-- --:--:-- --:--:-- 107k {} ``` You get an empty response, that's because in order to support the hivemind format and route queries to the correct SQL function, there's a router/preprocessor, (you can read about it here: https://gitlab.syncad.com/hive/hivemind/-/blob/develop/hive/db/sql_scripts/postgrest/home.sql So you should instead call http://localhost:8080**/rpc/home** : ``` curl 'http://localhost:8080/rpc/home' -X POST -H 'Content-Type: application/json' --data-raw '{"id":6,"jsonrpc":"2.0","method":"bridge.account_notifications","params":{"account":"gtg","limit":1}}' | jq % Total % Received % Xferd Average Speed Time Time Time Current Dload Upload Total Spent Left Speed 100 381 0 280 100 101 2552 920 --:--:-- --:--:-- --:--:-- 3495 { "id": "6", "result": [ { "id": 2111608, "msg": "@liondani voted on your post ($0.27)", "url": "@gtg/re-chitty-a-balanced-voting-system-is-impossible-with-current-witness-s-payouts-20160915t141551322z", "date": "2016-09-15T18:55:30", "type": "vote", "score": 50 } ], "jsonrpc": "2.0" } ``` And voilà ! That's it for this little tutorial, hope you found it useful :)