install.md

Edit makefile of your project to run each psql $(pg_uri)/$(pg_db) < pubLib$(i).sql in the correct order. For example any pubLib02-*.sql must run before pubLib03-*.sql, but same order like pubLib03-admin.sql and pubLib03-json.sql can be run in any order — by convention we adopt the alphabetic order, so use the ls pubLib*.sql order.

RECOMMENDATION: to install (in a database) only the used modules.

RATIONALE: each database in each project can use a different set of pubLib functions. It is good practice to avoid "namespace pollution" with unused functions, so, the pubLib modularization makes a reasonable partition over this "pollution risk".

Dependencies

The SQL source code, at /src, files in the form pubLib*.sql have its * part as labels in the folowing dependency diagram:

graph LR
    x3["03-{admin,json}"]
    x5h["05hcode-{distrib,encdec}"]
    x5gis["05pgis-{extraSRID,geohash,misc}"]
    00-general --> 01-array & 01py-string
    01-array --> 02-string --> x3 --> 04-aggregate
    04-aggregate --> x5h & x5gis & 05xml-general
    x5gis & x5h --> 06pgis-geoJSON

Loading

Installing all library

Use make instAll to install all scripts, or install at database db a specific one like array,

psql postgres://postgres@localhost/db < pubLib01-array.sql

SQL scripts of /src with pgis (file named pubLib{depLevel}pgis-{name}.sql) or py (named pubLib{depLevel}py-{name}.sql) are exceptions, need some more installing procedures.

py dependences

All Python functions are standard Python3, and need PLpython3 untrusted module:

sudo apt install  postgresql-plpython3-13

Special modules need some install procedures by pip, but to be visible at postgreSQL need sudo pip and unmask for permission. Example with Mustasche template Chevron module:

# pegadinha no path de instalação do módulo, ver https://stackoverflow.com/a/69811157/287948
sudo su
cd ~
umask 022
sudo pip3 install  chevron

pgis dependences

Use apt or instructions at https://postgis.net/install/