Bamboo Automation

This set of tools should allow to automate some of the task you want to do on Bamboo. It uses the rest api whenever possible and the web ui otherwise.

Install

Checkout the source from git

Structure

lib/requests.py
lib/bamboo_commands.py
lib/manipulate_bamboo_json.py

lib/requests.py handles the http requests
lib/bamboo_commands.py provides bamboo actions like add_job_requirement()
lib/manipulate_bamboo_json.py retrieve values from the bamboo results
main dir holds example files. that will work on a local dev bamboo, but be notoriously out of date

Examples

This section lists the usable commands.

They assume that you have imported the libraries like this

from lib.bamboo_commands import *
from lib.manipulate_bamboo_json import *

Create Connection

Either use username and password

conn = authenticate('http://localhost:6990', 'admin', 'admin', '/bamboo')

or take a JSESSIONID cookie from a file in the Mozilla format

conn = external_authenticate('http://localhost:6990', 'cookies.txt', '/bamboo')

(there is a firefox extension to export cookies)

The last parameter specifies a directory prefix, where bamboo can be found. If you don't need a prefix, don't specify it or make it empty ('').

List Projects

Lists all projects

project_dict = get_projects(conn)

You can use the expand option with the valid parameters from [bamboo's REST API definition] (https://developer.atlassian.com/display/BAMBOODEV/Bamboo+REST+Resources#BambooRESTResources-ProjectService "Bamboo REST API Projects")

project_dict = get_projects(conn, expand="projects.project.plans)

List Plans

List all plans

plan_dict = get_plans(conn)

project_dict = get_projects(conn, expand="plans.plan)

Get Project and Plan Keys

Get the keys from the bamboo json output. Plan keys include the project key in a projectkey-plankey schema, so you can access the plan with the plan key only. The keys are unique string representations for a project or plan.

project_keys = get_project_keys(project_dict)
plan_keys = get_plan_keys(plan_dict)

Manipulate Variables

Get the plan key first

plan_key = 'PROJECTKEY-PLANKEY'

Add a Variable

This will fail if the variable already exists

res = add_plan_variable(conn, plan_key, 'varkey', varvalue')

Modify a Variable

This will only work if the variable already exists

res = mod_plan_variable(conn, plan_key, 'varkey', 'varvalue')

Add/Modify a Variable

Tries to add the variable and modifies if it exists

res = add_mod_plan_variable(conn, plan_key, 'varkey', 'varvalue')

Delete Variables

You can delete one variable specified by key or delete all at once

res = delete_plan_variable(conn, plan_key, 'varkey')
res = delete_plan_all_variables(conn, plan_key)

Get Job ID

Get the ID of the plan's jobs. For jobs the key is the name presented on the website, the id the unique internal string representation. The result is different from get _ projects() or get _ plans(). It returns a dict of tuples.

job_key = 'Build'
job_dict = get_jobs(conn, plan_key)
job_id = job_dict[job_key][0]

Manipulate Requirements

Requirements are job-bound, so we need the job ID

job_id = 'PROJECTKEY-PLANKEY-JOB1'

Add a Requirement

This will fail if the requirement already exists. If you add a requirement that is already in the system, the parameter req _ exists must be True.

res = add_job_requirement(conn, job_id, 'reqkey', 'reqvalue', req_exists=False)

Delete Requirements

You can delete one requirement specified by key or delete all at once

res = delete_job_requirement(conn, job_id, 'reqkey')
res = delete_job_all_requirements(conn, job_id)

Manipulating Tasks

Tasks are job-bound, so we need the job ID. We also need the task key and the task parameters. This example shows the sourcecode checkout task. Fields that are not needed, like userDescription, can be left empty ("").

job_id = 'PROJECTKEY-PLANKEY-JOB1'


task_key = "com.atlassian.bamboo.plugins.vcs:task.vcs.checkout"
task_params = {
    "checkBoxFields": "cleanCheckout",
    "checkoutDir_0": "<where_to_checkout>",
    "selectFields": "selectedRepository_0",
    "selectedRepository_0": "defaultRepository",
    "userDescription": "<my_task_description>"
    }

Add a Task

A task can be added multiple times without problems

res = add_job_task(conn, job_id, task_key, task_params)

Move a Task

In bamboo you can move a task precisely. This function, however, only puts tasks into the final stage or back so far. This function needs the task id, which can be best retrieved with the task description.

task_description = "<my_task_description>"
task_dict = get_tasks(conn, job_id)
for task_id, (task_key, description, edit_link, del_link) in job_tasks.iteritems():
  if description == task_description:
    res = move_job_task(conn, job_id, task_id, finalising=True)

Delete Task

This function needs the task id, which can be best retrieved with the task description.

task_description = "<my_task_description>"
task_dict = get_tasks(conn, job_id)
for task_id, (task_key, description, edit_link, del_link) in job_tasks.iteritems():
  if description == task_description:
    res = delete_job_task(conn, job_id, task_id)

Change Permissions

You can change a permission by specifying it as a four-tuple: (usertype, username, permissiontype, value). All other values remain unchanged. This function is not very efficient for multiple changes, since it requests a permission list from bamboo on every invocation.

The usertype takes values 'user', 'group' or 'other.
The username is the short username in bamboo.
The permissiontype is 'read', 'write', 'build', 'clone', 'admin', or 'all' to change all permissions for this user.
The value is either True or False.

change_plan_permission(conn,
                       plan_key,
                       ('user','admin','all',True))
change_plan_permission(conn,
                       plan_key,
                       ('group','devel','build',True))
change_plan_permission(conn,
                       plan_key,
                       ('other','Anonymous Users','all',False))

Common Operations

These are hints how to do common operations on results. They might not be optimal solutions.

Delete a Number of Specific Tasks

tasks_to_delete = [
    'my install task',
    'my build task'
    ]
for task_id, (task_key, description, _, _) in job_tasks.iteritems():
  if description in tasks_to_delete:
    res = delete_job_task(conn, install_job_id, task_id)

Add Multiple Requirements

requirements = {
  "system.arch": "x86_64",
  "system.osfamily": "RedHat"
  }
for req_key, req_value in requirements.iteritems():
  res = add_job_requirement(conn, install_job_id, req_key, req_value)

How To Get the Task Parameters

Take Firefox with the firebug plugin and open the network tab. Make sure that you have 'persist' set, so connection details stay after page reloads. Add a task and look at the POST parameters. You don't have to specify all of them, the common ones are set by the function add _ job _ task(). If you're unsure, check the source.

Add an Inline Script Task

scriptBody = """<the whole script content surrounded by three quotes for multiline strings>"""

cp_task_key = "com.atlassian.bamboo.plugins.scripttask:task.builder.script"
cp_task_params = {
    "argument": "",
    "environmentVariables": "",
    "script": "",
    "scriptBody": scriptBody,
    "scriptLocation": "INLINE",
    "selectFields": "scriptLocation",
    "userDescription": "<description>",
    "workingSubDirectory": ""
    }

Get Plans from one Project

project = 'MYPROJ'
print 'PROJECT = %s' % project

plan_list = get_plans(conn, 'plans.plan.actions')
plan_keys = get_plan_keys(plan_list)
print 'PLAN KEYS = %s' % project

myproj_plan_keys = filter(lambda s: re.match(project,s) != None, plan_keys)
print 'PLAN KEYS IN PROJECT %s = %s' % (project, myproj_plan_keys)

yanxiaosong / bamboo-automate Goto Github PK

bamboo-automate's Introduction