A Plotly-Dash Python web application, developed by Achieve Intelligence, that generates tables and visualisations from pre-existing study milestone data for analysis and migration scheduling.
Achieve Intelligence offers a fully validated version of this tool. For enquires regarding this validated version, please contact Achieve Intelligence via email at: info@achieveintelligence.com
By default, this application will attempt to load the example csv data provided with this repository. Please replace this example data with your own csv or xlsx file and update the other configuration options found in config.py as required.
Run app.py using a virtual environment containing the packages found in the requirements.txt file. By default this launches the application on localhost port 8050.
The host or port can be changed setting the following environment variables as described by the app.run_server function:
host: str = os.getenv("HOST", "127.0.0.1"),
port: int = os.getenv("PORT", "8050"),
proxy: str = os.getenv("DASH_PROXY", None)
See the following guides for deployment information suitable for this application:
Azure - Python web app quickstart
Please see the domino documentation for more information on dash app publication with Domino.
The application comprises two main areas:
- Sidebar: Contains input fields and controls for filtering the data.
- Main Content Area: Displays the visualisations and tables generated based on the selected filters.
On startup, this application will load source data and populate the main study activity overview of milestone activity following the configuration specified in config.py.
Coloured blocks of activity are shown in the visualisations that represent the milestone dates from your source data plus/minus the offsets specified on the collapsable menu to the left of the screen. The timeframe and milestones located on this menu can be modified at any time.
The side menu also includes a "Hide Milestones Outside Timeframe" option which can be used to temporarily override the timeframe and return to a complete activity overview.
Below this main overall activity view, additional search functionality is provided to generate study or compound specific timeline views as required.
All the graphs provide additional information on milestone activity when hovering alongside several different tools and utilities in a menu located in the top right corner. The legend on these graphs is also interactive, allowing you to quickly toggle the visibility of specific milestones on click.
In version 2 of this tool, a Migration Table section and additional custom filters were added. These custom filters
can be applied from the sidebar to any columns and affect both the graphs and the migration table. More information
about the migration table can be found here.
Use the sidebar controls to configure the visualisations and migration table as follows:
- Timeframe: Specify the desired timeframe for the visualisations or migration analysis.
- Hide Milestones Outside Timeframe: This option can be used to override the timeframe in visualisations but does not affect the migration table.
- Milestone Offsets: Select the milestones to be included in the analysis by checking the corresponding checkboxes in the sidebar. The selected milestones will be used when generating the visualisations and migration table.
- Custom Filters: Apply additional filters by selecting column labels and their corresponding values. The available column labels depend on the dataset being used.
Click the "Update Global Configuration" button to update the graphs and tables based on the selected filters. The data will be processed, and the visualisations will be generated accordingly.
The migration table provides insights into study migrations within defined time periods. Using the Global Filters, frequency, studies transferred per (day/week) and length of period (days/weeks) the application will attempt to find optimal migration dates for studies within the timeframe.
Studies are sampled for a migration date in order of their perceived migration difficulty (derived from the total length of inactivity within the timeframe and the number of periods of inactivity).
The following controls are available for the migration table:
This expansion panel allows you to calculate a rough estimate for studies transferred per day or week based on the following numbers:
- The Working hours in a day input allows you to specify on average how many hours per day you are able to transfer studies.
- The Working days in a week input allows you to estimate the number of working days in a week you will be working on transferring studies.
- The Average size of study input is the average size of all the studies in your dataset. This can be given in KB, MB, GB or TB.
- The Average transfer rate input is the average speed at which the studies can be migrated. This can be given in KB/s, MB, or GB/s.
This rough estimate can be overwritten by supplying a value to the Studies transferred per day input below the study transfer calculator expansion panel.
The frequency control allows you to choose between migration periods measured in days or weeks. For weekly migration periods, you can set the number of days in a week using the day length control.
The grouping column dropdown allows you to select a column from the dataset for grouping studies. Choose a suitable column to group studies and analyse migrations based on that grouping.
Adjust the period length to define the duration of each migration period. The app will then calculate the number of periods which can fit inside the provided timeframe.
After configuring the migration parameters, click the "Generate Table" button to generate the migration table. The table will display study migrations and periods of inactivity based on the selected parameters.
Clicking this button will download the information used to produce the table into two csv files:
- migration_configuration_export_(timestamp).csv which contains information on the configuration used when the migration table is generated.
- migration_export_(timestamp).csv which provides a complete overview about each study's migration.
With the release of version 2, there are no new major features currently planned for the study migration tool. If you have any suggestions, please create a GitHub issue labelled enhancement.
We encourage you to contribute to the Study Migration Tool!
This application is built using Dash, which is a Python framework created by Plotly that uses Flask, Plotly.js and React under the hood. If you are unfamiliar with this framework, please see the Dash documentation.
To start contributing or report an issue, see our Contribution Guidelines.
For support, please contact the OpenSCE community.
The Study Migration Tool is released under the GNU General Public Licence.