Skip to content

Latest commit

 

History

History
35 lines (22 loc) · 1.38 KB

README.md

File metadata and controls

35 lines (22 loc) · 1.38 KB

As its name suggests, the CSV2LDIF python script allows one to create a LDIF file from a CSV.

Pre-Requisites

Python should be installed and passlib available

2 CSV files are required:

  • the users file whose path is defined as "USERS_FILE" variable
  • the orgs file whose path is defined as "ORGS_FILE" variable

The CSV files are expected:

  • to be defined in "USERS_FILE" and "ORGS_FILE" variables
  • to use commas as separators
  • to have quotes (double) around fields
  • to expose LDAP user fields in this order: employeeNumber, uid, userPassword, mail, givenName, sn, o, title, telephoneNumber, postalAddress, groups (cf. fields name in users.csv)
  • to expose LDAP orgs fields in this order: id, ou (short name), o (name), businessCategory, postalAddress, description, registered (cf. fields name in orgs.csv)

Note that:

  • uid, userPassword, mail & groups fields are mandatory, the others can be left blank (but not skipped).
  • roles names should be separated by semicolons

Usage

python CSV2LDIF.py > my.ldif

This script has been tested with Python 3.6. It need the "base64" module to work.

Note that the LDIF file will be generated for the dc=georchestra,dc=org root dn. To change this, feel free to edit the script.

The generated LDIF file is meant to be inserted in the LDAP right after the georchestra-bootstrap.ldif file, as it creates the full tree, including the root dn.