Skip to content

General Description

Liran Funaro edited this page May 11, 2017 · 4 revisions

The Graph Generation System was designed to provide an interface for graph generation based on experiment data, using a client-server web interface.

Using this system, the user can connect to a website, select a specific data file, graph model and parameters, and plot a graph in the browser.

The graphs are generated using the Bokeh library in python and are interactive - allowing zoom pan and save operations.

The system also allows to save/load/delete presets plots for different data files, as well as an option to define your own plugin.

Features

An interface for graph generation from given data file.

  • Interactive file explorer for easier experiment selection
  • Load preset: allows the user to load a preset per data file, selecting a preset from a list and generating a graph from it in a single click.
  • Create preset: allows the user to create a new graph preset from available graph models (line, step, heatmap etc..).
  • Save preset: allows the user to save the selected parameters as a preset for other users to use.
  • Plugin system: allow the user to define their own graph model by writing their own plugin in python.
  • Select-by-value: allows the user to define fields as "filter by value", which upon selection will display an additional input field to select by value. Select-by-value Fields can define a single or multiple selections.
  • Export: Generated graph's data is also saved as a JSON (under the "exports" sub-directory) and the plot can be saved as an image (PNG) upon request.
  • Zoom/Pan: Generated graphs are interactive in the browser, allowing zoom and pan.
  • Portability: Database files (SQLite) holds the models and presets, and can be moved between machines to retain models/presets information.