Thanks @jreyesr for this new plugin. Great work 🎉 !!

The basic structure looks good so far. While using the plugin, we did come up with a few suggestions based on our best practices:

postgres.spc

• Could you please update the postgres.spc file to follow the format of the namecheap plugin
• We should mention the config arguments however commented out and the values should be in the format of the actual values (however incorrect)
• For instance: There are only a few formatting issues that have been resolved in this example:

connection "postgres" {
  plugin = "jreyesr/postgres"

    # A connection string (https://pkg.go.dev/github.com/jackc/pgx/v5#hdr-Establishing_a_Connection), in the form that is 
    expected by the pgx package. Required. 
    # Can also be set with the `DATABASE_URL` environment variable.
    # connection_string = "postgres://username:password@localhost:5432/database_name"
  
    # The remote DB's schema that this plugin will expose. If you leave this unset, it'll default to `public`.
    # schema = "public"
  
    # List of tables that will be exposed from the remote DB. No dynamic tables will be created if this arg is empty or not set.
    # Wildcard-based searches are supported.
    # For example:
    #  - "*" will expose every table in the remote DB
    #  - "auth-*" will expose tables whose names start with "auth-"
    #  - "users" will only expose the specific table "users"
    # You can have several items (for example, ["auth-*", "users"] will expose 
    # all the tables that start with "auth-", PLUS the table "users")
    # Defaults to all tables
    # tables_to_expose = ["*"]
  }

• Is the environment variable name correct?

docs/tables

• The table name should be represented like # Table: raw.
• The formatting of the query should be updated using https://www.freeformatter.com/sql-formatter.html#before-output (2 indent spaces)
• The queries are not correctly formatted. They should end with a ;. Please refer to the following query:

select 
  name,
  id 
from
  aws_s3_bucket
where 
  region = 'us-east-1';

• We generally do not include the example query output in the docs, is there a particular reason for keeping them here?
• The example queries in the doc look a bit light. Could you please add 2 to 3 more example queries wherever possible?

docs/index.md

• The description should be updated to "description: Steampipe plugin for proxying queries to plain Postgres databases."
• The og_description should be updated to "Query any Postgres table from Steampipe with SQL! Open source CLI. No DB required."
• The formatting of the query should be updated using https://www.freeformatter.com/sql-formatter.html#before-output (2 indent spaces)
• The query is not correctly formatted. They should end with a ;.
• We generally do not include the example query output in the docs, is there a particular reason for keeping them here?
• One example query is enough in the index.md file we do not have to keep multiple examples here.
• The documentation and install sections are overlapping with the query section, I believe there are some issues with the "``" usage, please check that.
• Also please make sure that the Configuration section is in sync with the changes suggested for the config/postgres.spc file.
• We need a slight change in the way we include details of the environment variable, can you please change the section after the connection as follows:
  Alternatively, you can also use the standard Postgres environment variables to obtain credentials only if the other argument (connection_string) is not specified in the connection:

  export DATABASE_URL=postgres://username:password@localhost:5432/database_name

• Is the environment variable name correct?
• Under Get Involved we should include the links in - rather than * for instance

- Open source: https://github.com/jreyesr/steampipe-plugin-postgres
- Community: [Join #steampipe on Slack →](https://turbot.com/community/join)

• postgres
  • Please rename the config.go to connection_config.go
  • plugin.go
    • The function has some extra unnecessary spaces we can restructure it as follows

     func Plugin(ctx context.Context) *plugin.Plugin {
         p := &plugin.Plugin{
             Name: "steampipe-plugin-postgres",
             DefaultTransform: transform.FromGo().NullIfZero(),
             ConnectionConfigSchema: &plugin.ConnectionConfigSchema{
       	      NewInstance: ConfigInstance,
       	      Schema:      ConfigSchema,
             },
             SchemaMode:   plugin.SchemaModeDynamic,
             TableMapFunc: PluginTables,
         }
         return p
     }
    

README

• We can update Install the plugin with [Steampipe](https://steampipe.io) to Download and install the latest Postgres plugin:
• Also please make sure that the Configuration section is in sync with the changes suggested for the config/postgres.spc file.
• Please include the details of authentication via environment variable using the EXPORT command
• Please include the following inder contribution section:

`help wanted` issues:

- [Steampipe](https://github.com/turbot/steampipe/labels/help%20wanted)
- [Postgres Plugin](https://github.com/jreyesr/steampipe-plugin-postgres/labels/help%20wanted)

go.mod

• We should recompile the plugin with steampipe-plugin-sdk v5.6.1 and GO version 1.21

• A couple of follow-up questions:

  • How are we handling the resource not found errors?
  • Do the APIs support pagination? I did not find any reference to it, but I want to make sure that we have checked that box before releasing it.
  • While working with the APIs did you encounter rate limit errors?
  • Please refer to https://steampipe.io/docs/develop/plugin-release-checklist#data-ingestion for more information

Please let us know if you have questions, happy to help 👍

Warning:
    failed to start plugin 'hub.steampipe.io/plugins/jreyesr/postgres@latest': failed to plugin initialise plugin 'steampipe-plugin-postgres': TableMapFunc 'PluginTables' had unhandled error: interface conversion: interface is nil, not hclog.LoggerUpdatedConnections: true

I figured this must be a connection string problem, but I'm not seeing it.

My .spc:

 connection "postgres" {
  plugin = "jreyesr/postgres"

  # Write a connection string, in the form that is expected by the pgx package:
  # https://pkg.go.dev/github.com/jackc/pgx/v5#hdr-Establishing_a_Connection
  # Required
  connection_string = "postgres://jon:jon@localhost:5432/postgres"

  # The remote DB's schema that this plugin will expose
  # If you leave this unset, it'll default to `public`
  # schema = "public"
}

This works: psql -d postgres -U jon -W

I instrumented GetConfig like so:

func GetConfig(connection *plugin.Connection) PostgresConfig {
	logger.Debug("GetConfig", "connection", connection, "connection.Config", connection.Config)
	if connection == nil || connection.Config == nil {
		return PostgresConfig{}
	}
	config, other := connection.Config.(PostgresConfig)
	logger.Debug("GetConfig returning", "config", config, "other", other)
	return config
}

The output:

2023-09-11T10:40:59.563-0700 [DEBUG] plugin: postgres.PluginTables
2023-09-11T10:40:59.563-0700 [DEBUG] plugin: GetConfig: connection="&{postgres %!v(PANIC=String method: runtime error: invalid memory address or nil pointer d
ereference)}" connection.Config="%!v(PANIC=String method: runtime error: invalid memory address or nil pointer dereference)"
2023-09-11T10:40:59.563-0700 [DEBUG] plugin: GetConfig returning: config="%!v(PANIC=String method: runtime error: invalid memory address or nil pointer dereference)" other=true

I feel like I'm missing something really dumb?

Per the Plugin Release Checklist,

If the plugin can generate dynamic tables, a configuration argument should allow users to specify which tables the plugin will generate. This configuration argument typically accepts a list of strings and should support filesystem glob patterns like in the CSV plugin.

If this configuration argument is not set or is explicitly empty, e.g., paths = [], then no dynamic tables should be generated.

Add a configuration param to ignore some tables, such as those that have non-handled columns.