Update the autohelp readme
Switch to RST. Only document the wrapper usage, since it handles all the environment setup and is easier to use than the python scripts. Add a requirements.txt file to ease the setup process and documentation. Change-Id: Icc4f9cb4dfcaae6df8cd2567cc9b8f46fda0600c
This commit is contained in:
Gauvain Pocentek
parent
d0e1a97df9
commit
f4f6841278
|
|
@ -1,191 +0,0 @@
|
|||
autogenerate_config_docs
|
||||
========================
|
||||
|
||||
Automatically generate configuration tables to document OpenStack.
|
||||
|
||||
|
||||
Dependencies: python-git (at least version 0.3.2 RC1), oslo-incubator.
|
||||
|
||||
Using the wrapper
|
||||
-----------------
|
||||
|
||||
The autohelp-wrapper script installs a virtual environment and all the needed
|
||||
dependencies, clones or update the projects and manuals repositories, then runs
|
||||
the autohelp.py script.
|
||||
|
||||
The workflow is:
|
||||
|
||||
$ ./autohelp-wrapper update
|
||||
$ vim sources/openstack-manuals/tools/autogenerate-config-flagmappings/*.flagmappings
|
||||
$ ./autohelp-wrapper docbook
|
||||
# check the results
|
||||
$ git commit -a
|
||||
|
||||
For details about the autohelp.py script and its manual setup, see the next
|
||||
sections.
|
||||
|
||||
Setting up your environment
|
||||
---------------------------
|
||||
|
||||
Note: This tool is best run in a fresh VM environment or a python virtual
|
||||
environment, as running it requires installing the dependencies of the
|
||||
particular OpenStack product you are working with. Installing all of that on
|
||||
your normal machine could leave you with a bunch of cruft!
|
||||
|
||||
Install git and python-pip:
|
||||
|
||||
$ sudo apt-get install git python-pip
|
||||
|
||||
Install oslo.config and GitPython:
|
||||
|
||||
$ sudo pip install oslo.config "GitPython>=0.3.2.RC1"
|
||||
|
||||
Install oslo-incubator from git:
|
||||
|
||||
$ git clone git://git.openstack.org/openstack/oslo-incubator
|
||||
$ cd oslo-incubator
|
||||
$ python setup.py build
|
||||
$ sudo python setup.py install
|
||||
|
||||
Check out the repository you are working with:
|
||||
|
||||
$ git clone https://git.openstack.org/openstack/nova.git
|
||||
|
||||
(This guide makes reference to a /repos directory, so you should
|
||||
record the directory you are using and replace as appropriate below.)
|
||||
|
||||
Check out the tool itself:
|
||||
|
||||
$ git clone https://git.openstack.org/openstack/openstack-doc-tools.git
|
||||
|
||||
Install the dependencies for the product you are working with:
|
||||
|
||||
$ sudo pip install -r nova/requirements.txt
|
||||
|
||||
Note 1: Occasionally, the requirements.txt file does not reference all
|
||||
dependencies needed to import all modules in a package. You may need
|
||||
to manually install some dependencies with pip or your distro's
|
||||
package manager. Always check the log for failed imports.
|
||||
|
||||
Note 2: Although this tool imports from your checkout, occasionally
|
||||
conflicts happen with installed files (e.g. a change from `foo/bar.py`
|
||||
to `foo/bar/__init__.py`). For best results, uninstall the old product
|
||||
and install from git using setup.py.
|
||||
|
||||
|
||||
Using the tool
|
||||
--------------
|
||||
|
||||
This tool is divided into three parts:
|
||||
|
||||
1. Extraction of flags names:
|
||||
|
||||
$ openstack-doc-tools/autogenerate_config_docs/autohelp.py create nova -i /repos/nova
|
||||
|
||||
You only need to use the `create` action to add a new product.
|
||||
Otherwise, use `update`:
|
||||
|
||||
$ openstack-doc-tools/autogenerate_config_docs/autohelp.py update nova -i /repos/nova
|
||||
|
||||
The `create` action will create a new `nova.flagmappings` file,
|
||||
possibly overriding an existing file. The `update` action will
|
||||
create a `nova.flagmappings.new` file, merging in group
|
||||
information from the existing `nova.flagmappings` file.
|
||||
|
||||
2. Grouping of flags
|
||||
|
||||
This is currently done manually, by using the flag name file and placing
|
||||
a category after a space:
|
||||
|
||||
$ head nova.flagmappings
|
||||
aggregate\_image\_properties\_isolation\_namespace scheduling
|
||||
aggregate\_image\_properties\_isolation\_separator scheduling
|
||||
allow\_instance\_snapshots policy
|
||||
allow\_migrate\_to\_same\_host policy
|
||||
allow\_resize\_to\_same\_host policy
|
||||
allow\_same\_net\_traffic network
|
||||
allowed\_direct\_url\_schemes glance
|
||||
...
|
||||
|
||||
3. Creation of docbook-formatted configuration table files:
|
||||
|
||||
Example:
|
||||
|
||||
$ openstack-doc-tools/autogenerate_config_docs/autohelp.py docbook nova -i /repos/nova
|
||||
|
||||
|
||||
Duplicate options
|
||||
-----------------
|
||||
|
||||
Sometimes the log will tell you there was a duplicate option name.
|
||||
This often happens if a parent module does `import *` from a child
|
||||
module. For example, `nova.db` does `from nova.db.api import *`.
|
||||
If one module name is a child of the other, the tool will use the
|
||||
option from the more specific module and tell you so. This is
|
||||
probably not a problem.
|
||||
|
||||
If the tool tells you it's chosing one at random, then something
|
||||
else is happening, and you should investigate. Otherwise, you may
|
||||
lose an option.
|
||||
|
||||
A worked example - updating the docs for Icehouse
|
||||
-------------------------------------------------
|
||||
|
||||
Update automatically generated tables - from scratch, using a virtual
|
||||
environment. Using a python virtual environment will avoid modifications and
|
||||
conflicts with python packages installed on the base system. The virtual
|
||||
environment can be destroyed by deleting the `venv` directory.
|
||||
|
||||
Note that the virtual environment must be activated for every new shell you
|
||||
start. This can be done by sourcing the `venv/bin/activate` file.
|
||||
|
||||
$ mkdir WORKDIR
|
||||
$ cd WORKDIR
|
||||
$ sudo apt-get update
|
||||
$ sudo apt-get install git python-virtualenv
|
||||
$ virtualenv venv
|
||||
$ source venv/bin/activate
|
||||
$ git clone git://git.openstack.org/openstack/openstack-doc-tools.git
|
||||
$ git clone git://git.openstack.org/openstack/openstack-manuals.git
|
||||
$ git clone git://git.openstack.org/openstack/oslo-incubator
|
||||
$ cd oslo-incubator
|
||||
$ python setup.py install
|
||||
$ cd ..
|
||||
$ pip install "GitPython>=0.3.2.RC1"
|
||||
$ cd openstack-manuals/tools/autogenerate-config-flagmappings
|
||||
|
||||
Now, download and install the projects. Installation is necessary to satisfy
|
||||
dependencies between projects. This will also install required dependencies.
|
||||
|
||||
$ mkdir -p sources && cd sources
|
||||
$ PROJECTS="ceilometer cinder glance heat neutron nova trove"
|
||||
$ for p in $PROJECTS; do git clone git://git.openstack.org/openstack/$p.git; done
|
||||
$ for p in $PROJECTS; do cd $p && python setup.py install && cd ..; done
|
||||
$ cd ..
|
||||
|
||||
Install some missing requirements:
|
||||
|
||||
$ pip install fixtures swift hp3parclient ryu pymongo bson pbr
|
||||
|
||||
Note that some dependencies will still be missing. They can't be installed using
|
||||
pip, they should be installed manually. Not installing those dependencies will
|
||||
not prevent the tool to work, but some configuration options might not be
|
||||
discovered.
|
||||
|
||||
Update the flag names:
|
||||
|
||||
$ for p in $PROJECTS; do
|
||||
> ../../../openstack-doc-tools/autogenerate_config_docs/autohelp.py -vvv update $p
|
||||
> done
|
||||
|
||||
At this point, search through the `*.flagmappings.new` files for anything
|
||||
labelled `Unknown` and fix it. Once that is done rename the `*.flagmappings.new`
|
||||
files and run `autohelp.py` with the `docbook` subcommand:
|
||||
|
||||
$ for p in $PROJECTS; do
|
||||
> mv $p.flagmappings.new $p.flagmappings
|
||||
> ../../../openstack-doc-tools/autogenerate_config_docs/autohelp.py -vvv docbook $p
|
||||
> done
|
||||
|
||||
to generate the XML files and move those into the appropriate part of
|
||||
the git repo.
|
||||
|
|
@ -0,0 +1,93 @@
|
|||
autogenerate_config_docs
|
||||
========================
|
||||
|
||||
Automatically generate configuration tables to document OpenStack.
|
||||
|
||||
Using the wrapper
|
||||
-----------------
|
||||
|
||||
``autohelp-wrapper`` is the recommended tool to generate the configuration
|
||||
tables. Don't bother using ``autohelp.py`` manually.
|
||||
|
||||
The ``autohelp-wrapper`` script installs a virtual environment and all the
|
||||
needed dependencies, clones or updates the projects and manuals repositories,
|
||||
then runs the ``autohelp.py`` script in the virtual environment.
|
||||
|
||||
New and updated flagmappings are generated in the ``openstack-manuals``
|
||||
repository (``tools/autogenerate-config-flagmappings/`` directory).
|
||||
|
||||
The workflow is:
|
||||
|
||||
.. code-block:: console
|
||||
|
||||
$ pip install -rrequirements.txt
|
||||
$ ./autohelp-wrapper update
|
||||
$ $EDITOR sources/openstack-manuals/tools/autogenerate-config-flagmappings/*.flagmappings
|
||||
$ ./autohelp-wrapper docbook
|
||||
$ # check the results in sources/openstack-manuals
|
||||
|
||||
This will generate the tables for all the known projects.
|
||||
|
||||
To generate the mappings and tables for a subset of projects, use the code
|
||||
names as arguments:
|
||||
|
||||
.. code-block:: console
|
||||
|
||||
$ ./autohelp-wrapper update cinder heat
|
||||
$ # edit the mappings files
|
||||
$ ./autohelp-wrapper docbook cinder heat
|
||||
|
||||
|
||||
Creating mappings for a new project
|
||||
-----------------------------------
|
||||
|
||||
Run the wrapper with the create subcommand:
|
||||
|
||||
.. code-block:: console
|
||||
|
||||
$ ./autohelp-wrapper create zaqar
|
||||
|
||||
|
||||
Flagmappings files
|
||||
------------------
|
||||
|
||||
The flagmappings files use the following format:
|
||||
|
||||
::
|
||||
|
||||
OPTION_SECTION/OPTION_NAME group1 [group2, ...]
|
||||
|
||||
Groups need to be defined manually to organize the configuration tables.
|
||||
|
||||
The group values can only contain alphanumeric characters, _ and - (they will
|
||||
be used as docbook IDs).
|
||||
|
||||
To make the table titles more user friendly, create or edit the PROJECT.headers
|
||||
file in the manuals repository. Each line of this file is of the form:
|
||||
|
||||
::
|
||||
|
||||
GROUP A Nice Title
|
||||
|
||||
Working with branches
|
||||
---------------------
|
||||
|
||||
``autohelp-wrapper`` works on the master branch by default, but you can tell it
|
||||
to work on another branch:
|
||||
|
||||
.. code-block:: console
|
||||
|
||||
$ ./autohelp-wrapper update -b stable/icehouse
|
||||
|
||||
.. note::
|
||||
The ``-b`` switch doesn't apply to the ``openstack-manuals`` repository
|
||||
which will be left untouched (no ``git branch``, no ``git update``).
|
||||
|
||||
|
||||
Updating swift options
|
||||
----------------------
|
||||
|
||||
Swift configuration tables are generated using the ``extract_swift_flags.py``
|
||||
script. This script doesn't use a mapping file, but organize the tables using
|
||||
the various configuration files and sections. Most of the options must be
|
||||
described manually at the moment.
|
||||
|
|
@ -0,0 +1,4 @@
|
|||
GitPython>=0.3.2.RC1
|
||||
lxml
|
||||
oslo.config
|
||||
oslo.i18n
|
||||
Loading…
Reference in New Issue