GraphDash
is a web-based dashboard built on graphs and their metadata.
For example, if you have:
$ ls default_graph_dir
graph.svg
graph.txt
The svg file is the graph, and must have metadata with YAML format:
$ cat default_graph_dir/graph.txt
name: graph.svg
family: Category 1
title: '*Real serious* graph'
You can put as many graphs as you want in the directory, then start the graph dashboard. You will get a web interface with search box, autocompletion and easy navigation.
$ GraphDash -r default_graph_dir
* Running on http://0.0.0.0:5555/ (Press CTRL+C to quit)
Clone and install (in user space):
git clone https://github.com/AmadeusITGroup/graphdash.git
cd graphdash
pip install --user .
Or use the Python package:
pip install --user graphdash
For user-space installation, make sure your $PATH
includes ~/.local/bin
.
$ GraphDash -r default_graph_dir
* Running on http://0.0.0.0:5555/ (Press CTRL+C to quit)
The dashboard can be configured with a YAML config file and the -c/--conf
option:
$ cat docs/example.conf
root: ../default_graph_dir
title: "Example of title ;)"
subtitle: "Example of subtitle"
$ GraphDash -c docs/example.conf
* Running on http://0.0.0.0:5555/ (Press CTRL+C to quit)
You can generate a template of configuration file:
$ GraphDash -C template.conf
If not already installed on your machine, install Gunicorn
:
pip install --user gunicorn # on Fedora you may need to install libffi-devel before
Since you can import the webapp through graphdash:app
, you can serve it with Gunicorn
:
gunicorn -b 0.0.0.0:8888 --pid server.pid graphdash:app &
The configuration file of the webapp can be set with the CONF
environment variable.
With Gunicorn
, you can pass environment variables to the workers with --env
:
gunicorn -b 0.0.0.0:8888 --pid server.pid --env CONF=docs/example.conf graphdash:app &
But you should not use these commands yourself, that is what GraphDashManage
is for!
GraphDashManage
is used to start
, stop
, restart
the
instances of Gunicorn
serving graphdash:app
. It needs a
configuration file in the current directory:
$ cat settings.sh
ALL_MODES=(
['prod']="docs/example.conf"
['test']="docs/example.conf"
)
ALL_PORTS=(
['prod']=1234
['test']=5678
)
Then you can manage multiple instances of GraphDash
using Gunicorn
with:
$ GraphDashManage start prod
[INFO] Listening at: http://0.0.0.0:1234
[INFO] Booting worker with pid: 30403
[INFO] Booting worker with pid: 30404
[INFO] Booting worker with pid: 30405
$ GraphDashManage start test
[INFO] Listening at: http://0.0.0.0:5678
...
You can generate a template of settings:
$ GraphDashManage template > template.sh # to be moved to settings.sh
Possible entries (everything is optional):
root
: the root directory of the graphsfamilies
: path to the families metadata file (optional)title
: the title of the webappsubtitle
: the subtitle of the webappplaceholder
: the default text in the search fieldheader
: an optional message at the top (markdown syntax)footer
: an optional message at the bottom (markdown syntax)showfamilynumbers
: a boolean to toggle family numbering (default is true)showgraphnumbers
: a boolean to toggle graph numbering (default is true)theme
: change css theme (default is dark)keep
: the proportion of common words kept for autocompletionlogfile
: change default log file of the webappraw
: when loading, look for all graphs and ignore txt metadataverbose
: a boolean indicating verbosity when loading applicationdebug
: debug mode (enable Grunt livereload, enable Flask debug mode)headless
: headless mode (only search is available, no page is rendered)port
: when launched with Flask development server only, port
Several attributes are supported:
name
: the path to the graphtitle
: title of the graph, recommended for display purposes (markdown syntax)family
: the subsection in which the graph isindex
: an optional list of keywords describing the graph (useful for search feature)text
: an optional description of the graph (markdown syntax)pretext
: an optional message appearing before the graph (markdown syntax)file
: optional path to the raw dataexport
: optional path to the exportable graph (for example, a PNG file)rank
: integer, optional value used to change graphs order (default uses titles)showtitle
: a boolean to toggle title display for the graph (default is false)labels
: a list of labels (like 'new') which will be rendered in the UI as colored circlesother
: other metadata not used by GraphDash, but may be needed by other things reading the txt files
Note that if the name
attribute is missing, the graph will not be
shown and the text will be displayed anyway, like a blog entry.
You may put a .FAMILIES.txt
file at the root of the graph directory.
This file may contain metadata for families. It should be a YAML list:
- family: chairs
rank : 0
- family: tables
rank : 1
text: This is a description
alias: This text will appear instead of "tables"
labels: new
Each element of the list should be a dict containing:
family
: the family consideredrank
: integer, optional value used to change families order (default uses family name)text
: an optional description of the family (markdown syntax)alias
: an optional name who may be longer than the one in the url (useful to build nice urls)labels
: a list of labels (like 'new') which will be rendered in the UI as colored circles
Available labels are "new", "update", "bugfix", "warning", "error", "ongoing", "obsolete". You may give other labels which will be rendered with defaults colors. For customization, you may specify your own labels with a dict syntax:
labels:
- name: newlabel
color: white
text_color: black
text: "NEW LABEL"
tooltip: null
If you wish to contribute, you need Grunt
to generate new css/js files
from sass/coffee source files.
npm install --no-bin-links # may need to repeat
grunt
Debugging can be made with source map files for browser supporting them
in their debugging tools. If not, the Gruntfile.js
enables an option
to generate non-minified assets.
grunt --dev
With the debug
mode enabled, Grunt will use the livereload mechanism
to reload the browser if any file has changed (and Flask debug mode will
reload the server as well).
GraphDash --debug & # or python -m graphdash
grunt watch
If you used Gunicorn
with a PID file, Grunt will automatically reload it
if any Python files change.
gunicorn -b 0.0.0.0:8888 --pid server.pid graphdash:app &
grunt watch
You can use tox
build packages and run tests.
tox