Update configuration docs

docs/configuration.md

@@ -32,27 +32,37 @@ Recap uses [pydantic-settings](https://docs.pydantic.dev/latest/usage/pydantic_s
 | :-- | :-- | :-- | :-: | :-: |
 | RECAP_CONFIG | TEXT | Path to the config file. | ~/.recap/config | YES |
 | RECAP_SECRETS | TEXT | Path to the secrets directory. | | NO |
-| RECAP_SYSTEMS__[SYSTEM] | TEXT | A system name/URL pair that tells Recap how to connect to a system. | | NO |
+| RECAP_URLS | TEXT| List of URLs in JSON format containing authorization credentials. | | NO |
+| RECAP_REGISTRY_STORAGE_URL | TEXT | Location to store registry schemas in an [`fsspec`-compatible URL format](https://filesystem-spec.readthedocs.io/en/latest/api.html#other-known-implementations). | | NO |
+| RECAP_REGISTRY_STORAGE_ARGS | TEXT | Additional attributes to supply when creating the registry's `fsspec` filesystem object. | | NO |
 
-{: .note }
-The `RECAP_SYSTEMS__` environment variable is a special case. It's used to configure systems. The `SYSTEM` part of the variable name is the name of the system. The value is the URL to connect to the system. For example, `RECAP_SYSTEMS__MY_PG=postgresql://user:pass@host:port/dbname` will configure a system called `my_pg` to connect to a PostgreSQL database.
+#### RECAP_URLS
+
+The `RECAP_URLS` environment variable is important. It allows you to define connection strings for URLs that contain login credentials like usernames and passwords or authentication tokens.
 
-## CLI
+For example, if you have a URL like `postgresql://user:pass@localhost:5432/testdb`, you can set the `RECAP_URLS` environment variable like so:
+
+```bash
+export RECAP_URLS='["postgresql://some_user:some_pass@localhost:5432/testdb"]'
+```
 
-Recap's CLI has [`add`](/docs/cli#add) and [`remove`](/docs/cli#remove) commands to add and remove systems from Recap's config file (`RECAP_CONFIG`). These commands are just shortcuts for editing the config file directly.
+All `ls` or `schema` calls for the URL will use the credentials defined in the environment variable.
 
-### Add a System
+You don't need to specify a full path for URL credentials. Recap will substitute in any username, password, or query parameters that are defined in the environment variable, so this command:
 
 ```bash
-recap add my_pg postgresql://user:pass@host:port/dbname
+curl http://localhost:8000/gateway/ls/postgresql://localhost:5432/testdb/public
 ```
 
-### Remove a System
+Would be translated to:
 
 ```bash
-recap remove my_pg
+curl http://localhost:8000/gateway/ls/postgresql://some_user:some_pass@localhost:5432/testdb/public
 ```
 
+{: .note }
+The URL matching is done on the scheme ("postgresql" in our example) and netloc ("localhost:5432" in our example) for a URL.
+
 ## Dotenv
 
 Recap supports [`.env`](https://github.com/motdotla/dotenv) files as well. Recap will look for a `.env` file in the current working directory and load it if it exists. If not, it'll recurse upwards until it finds one.
@@ -71,20 +81,10 @@ See [pydantic-settings](https://docs.pydantic.dev/latest/usage/pydantic_settings
 
 Recap supports a secrets directory (`RECAP_SECRETS`) that contains files with secrets. Each file's name corrseponds to its environment variable name, and each file's contents are the environment variable's value.
 
-For example, if Recap is configured to use `/path/to/secrets` as its secrets directory, and `/path/to/secrets/recap_systems__my_pg` contains `postgresql://localhost:5432/testdb`, then Recap will see `RECAP_SYSTEMS__MY_PG`'s value as `postgresql://localhost:5432/testdb`.
-
-{: .warning }
-> If you want to set more than one system, you'll need to use the secret filename `recap_systems`. The contents should be a JSON-encoded map from system name to system URL. For example:
-> 
-> ```json
-> {
->   "my_bq": "bigquery://",
->   "my_pg": "postgresql://localhost:5432/testdb"
-> }
-> ```
+For example, if Recap is configured to use `/path/to/secrets` as its secrets directory, and `/path/to/secrets/recap_urls` contains `["postgresql://user:passlocalhost:5432/testdb"]`, then Recap will use the credentials defined in the file for any `ls` or `schema` calls for the URL.
 
 See [pydantic-settings](https://docs.pydantic.dev/latest/usage/pydantic_settings#secrets)'s secrets section for more information.
 
 ### Docker
 
-If you're deploying Recap's [gateway](/docs/gateway) in Docker, see the [Gateway Docker](/docs/gateway#docker) documentation and [pydantic-settings](https://docs.pydantic.dev/latest/usage/pydantic_settings/#use-case-docker-secrets) for Docker secrets instructions.
+If you're deploying Recap's server in Docker, see the [gateway](/docs/gateway#docker) and [registry](/docs/registry#docker) documentation and [pydantic-settings](https://docs.pydantic.dev/latest/usage/pydantic_settings/#use-case-docker-secrets) for Docker secrets instructions.

docs/gateway.md

@@ -15,7 +15,7 @@ Running Recap's stateless HTTP/JSON gateway server.
 
 ## Overview
 
-Recap comes with a stateless HTTP/JSON server that you can use to query systems and schemas. The server doesn't require any database or storage. It's a simple gateway that connects to systems and returns Recap schemas.
+Recap's HTTP/JSON server comes with a gateway API that you can use to query systems and schemas. The server doesn't require any database or storage. It's simply a gateway that connects to systems and returns Recap schemas.
 
 ## Usage
 
@@ -30,36 +30,34 @@ See the [CLI](/docs/cli#serve) documentation for additional `serve` parameters a
 {: .note }
 The gateway server is a simple [FastAPI](https://fastapi.tiangolo.com/) application. The `serve` command starts the FastAPI application using [uvicorn](https://www.uvicorn.org/). You may choose any ASGI-compatible server. See [FastAPI's documentation](https://fastapi.tiangolo.com/deployment/manually/) for more information.
 
-## Configuration
-
-You must define Recap "system" connections as environment variables before starting the gateway server. Recap uses these environment variables to connect to systems.
-
-Here's an example for a PostgreSQL system:
+You can query a system like so:
 
 ```bash
-export MY_PG=postgresql://user:pass@host:port/dbname
+curl http://localhost:8000/gateway/ls/postgresql://user:pass@localhost:5432/testdb
 ```
 
-Or you can use Recap's CLI:
+And read a schema:
 
 ```bash
-recap add my_pg postgresql://user:pass@host:port/dbname
+curl http://localhost:8000/gateway/schema/postgresql://user:pass@localhost:5432/testdb/public/test_types
 ```
 
-After running `recap serve`, you can query the system using the `my_pg` name:
+## Configuration
+
+The gateway will work out of the box without any configuration. Any URL with a compatible [client integration](/docs/integrations/) will work. However, you can configure Recap to save access credentials in an environment variable using the `RECAP_URLS` environment variable.
 
-```bash
-curl http://localhost:8000/ls/my_pg
-```
+- `RECAP_URLS`: A JSON-encoded list of URLs to connect to.
+
+See the [configuration](/docs/configuration) for more information on the `RECAP_URLS` environment variable.
 
 You can also set environment variables in .env files or secrets files. See Recap's [configuration](/docs/configuration) documentation for more information.
 
 ## API
 
 Recap's gateway server exposes a very simple HTTP/JSON API with two endpoints:
 
-* `/ls/[path*]` - List children in a system path.
-* `/schema/[path*]` - Get a schema for a system path.
+* `/gateway/ls/[url]` - List children of a URL.
+* `/gateway/schema/[url]` - Get a schema for a URL.
 
 ### OpanAPI
 
@@ -86,7 +84,7 @@ You can run the server using the `docker run` command:
 ```bash
 docker run \
     -p 8000:8000 \
-    -e "RECAP_SYSTEMS__PG=postgresql://user:pass@localhost:5432/testdb" \
+    -e "RECAP_URLS='[\"postgresql://user:pass@pg:5432/testdb\"]'" \
     ghcr.io/recap-build/recap:latest
 ```