Gramex 1.53 release notes

Natural Language Generation

Gramex has a library and interface for Natural Language Generation. Admins can upload datasets, describe them in English, and Gramex automatically creates a template out of it. These templates are dynamic (based on data) and can be embedded in dashboards

The feature is in beta, and the documentation is not available yet. But you can preview it on the admin page.

Gramextest makes testing easier

Gramextest has several new features:

• skip: true lets you skip tests until you run skip: false
• debug: true pauses on every step until you run debug: false
• python: cmd runs a Python command. This can access global variables set in script:
• script: cmd runs a JavaScript command. This can access globals variaables set in python:
• print: selector prints the HTML contents of a selector
• mark: group groups tests together into groups
• browsers: supports a mobile: option that lets you specify mobile simulation

A blank tests/gramextest.yaml is now included in gramex init.

g1 features fuzzy search

g1.fuzzysearch lets you match strings approximately. This makes search-as-you-type much easier.

g1 is upgraded to 0.16.1. This improves $.template and the link: attribute in FormHandler columns.

Vega is upgraded to 5.0.

Templates support UI modules

Tornado supports a feature called UI Modules. This lets you import template files like calling functions.

FileHandler templates now support this feature. You can use:

This is the parent template. {% module Template('embedded.html', var1=...,
var2=...) %} -- this template is embedded here.

Statistics

The Gramex code base has:

• 18,345 lines of Python
• 3,574 lines JavaScript
• 10,354 lines of test code
• 81% test coverage

How to upgrade

To upgrade Gramex, run:

pip install --upgrade gramex
pip install --upgrade gramexenterprise    # If you use DBAuth, LDAPAuth, etc.
gramex setup --all