martin-config

A python tool to generate config.yaml for martin vector tiles server from PostGIS database leveraging asyncpg library.
martin server uses a YAML config file to specify what tables/functions from the PostGIS DB will be published.

• Features
• Requirements
• Installation
• Usage

##Features

martin-config uses asyncpg to interact with the Postgres server and supports following functionality:

• creating config for table sources
• creating config for function sources
• creating config for the general section
• operate in one, multiple or all schemas
• selective publishing of tables through table comments
• selective publishing of columns through column comments
• uploading the config file to an Azure File Share using SAS authentication

Its strongest feature is the ability to select only the mark tables and table columns. This is achieved by inserting the

publish=True/False.

into the comment of a given table/column. The folowing rules apply

• if a table does not have this string in the comments it is considered unmarked and unpublishable
• if a has this string in comments its value dictates if the table is publishable (True) or unpublishable (False)
• same logic applied to columns with the difference that unmarked columns are publishable by default

This allows users to mark/set the tables/columns selectively and use martin-config to generate a conforming config file.

Requirements

asyncpg, yaml, dotenv and optionally azure-file-share

Install

1. Create a pipenv based venv

python3 -m pip install pipenv
pipenv --python 3

2. install martin-config

pipenv run pip install martin-config

3. optionally install azure functionality to be able to upload the cfg to AFS

pipenv run pip install martin-config[azure]

Usage

the installer creates a command line script martincfg

with pipenv it can be invoked like this:

pipenv run martincfg

pipenv run martincfg
Loading .env environment variables...
usage: martincfg [-h] [-s DATABASE_SCHEMA [DATABASE_SCHEMA ...]] [-o OUT_CFG_FILE] [-e ENV_FILE] [-d] [-sfs]

Create a config file for martin vector tile server

optional arguments:
  -h, --help            show this help message and exit
  -s DATABASE_SCHEMA [DATABASE_SCHEMA ...], --database-schema DATABASE_SCHEMA [DATABASE_SCHEMA ...]
                        A list of schema names. If no schema is specified all schemas are used.
  -o OUT_CFG_FILE, --out-cfg-file OUT_CFG_FILE
                        Full path to the config file to be created. If not supplied the YAML fill be dumped to stdout
  -e ENV_FILE, --env-file ENV_FILE
                        Load environmental variables from .env file
  -d, --debug           Set log level to debug
  -sfs, --skip-function-sources
                        Do not create config for function sources

Environmental variables

The arguments are self-explanatory. However, environmental variables are also required to be available at runtime, specifically:

 POSTGRES_DSN=postgresql://[user[:password]@][netloc][:port][/dbname][?param1=value1&...]
 
 AZURE_FILESHARE_SASURL=https://[storageaccount].file.core.windows.net/[sharename]?[SAS_QUERY_STRING]

Tips:

• If no schema is specified all schemas are used.
• multiple schemas can be specified one after another public myschema or public,myschema
• .env file containing the above environmental variable can be used to the same effect as defining env. variables

For improving the performance of the script, please create statistics for all tables by the following SQL.

-- analyze all tables
ANALYZE;

-- analyze specific table
ANALYZE schema.tablename;