Flexible and lightweight tool for generating HTML-based electronic lab notebooks
Labnote is a flexible and lightweight tool for generating HTML-based electronic lab notebooks.
Rather than attempting to provide a unified tool for creating and sharing lab notebook entries, Labnote simply ties together existing documents and analyses outputs and builds and creates an HTML index of these resources.
In short, it helps you go from something like this:
├── animal_behavior │ └── molothrus │ └── README.html ├── barnacles │ ├── cirripede-morphology │ │ └── README.html │ └── cirripede-taxonomy │ └── README.html └── finches ├── finch-beak-size-comparison │ └── beak_size.py ├── finch-foraging-strategies │ └── foraging-strategies.py └── natural-selection └── thoughts.txt
To something like this:
Labnote works by scanning a set of one or more directories for files matching a pattern that you specify as pertaining to notebook entries (e.g. a single log, script, or document describing some particular project or analysis.) It then constructs an HTML table of contents pointing to each of the matching files. By default, results are sorted by last-modified date. Categories can be defined and used to separate entries relating to different topics.
In order to support as many different work styles as possible, labnote tries and make as few assumptions as possible about how your files are organized, and provides configuration options allowing for a wide range of directory structures and file types.
Finally, labnote is designed to be extensible. While currently there is only a single no-frills theme, the jinga2 templating system used by Labnote makes it trivial to create themes.
To use labnote, you must have a recent version of Python (>=3.3)) available on your machine.
Additionally, labnote requires the following Python libraries:
If you are using pip to install labnote, all of the required dependencies should be automatically installed for you.
Labnote is currently aimed at supporting Windows, Linux, and OS X setups.
To install labnote using pip, run:
pip install labnote
To generate the example notebook, cd to the labnote source directory and run:
labnote -c example/example.config.yml \ -i example/research/*/* \ -o example/research/index.html
A file named index.html should be outputted to the example/ directory and should look something like what is shown in the screenshot above.
Labnote can be easily automated using Cron. For example, to have labnote regenerate your lab notebook once a day, run crontab -e to edit your user-level cron jobs, and add:
If you have created a user configuration for labnote in $HOME/.config/labnote/config.yml, then you are all set. Otherwise simply add whatever options you would use when calling Labnote from the command-line to the cronjob, e.g.:
@daily labnote -c /path/to/config.yml
For more information on how to create and customize cron jobs on your system, see the Ubuntu Cron Tutorial.
Labnote settings can be specified either via the command-line at run-time (e.g. labnote -i /some/path/* -o /output/dir), or using a YAML config file, or both.
By default, Labnote looks for a file named config.yml located in $HOME/.config/labnote/. If this file exists, then it will be used used to configure Labnote’s behavior.
The configuration file should look something like:
--- # General information title: Lab Notebook author: Your Name email: email@example.com # Notebook contents input_dirs: - /home/user/Dropbox/research/201[2-5]/* - /home/user/Dropbox/research/2016/*/* output_file: /home/user/Dropbox/research/index.html include_files: ['*.html', '*.py', '*.ipynb', 'README.*'] # Research categories categories: 'Sequence Analysis': ['seq', 'dna', 'rna'] 'Differential Expression': ['dea', 'differential-expression'] 'Network Analysis': ['network'] 'Visualization': ['viz']
The main settings that are important to define are:
You can also point to a config file located in a different location using the -c option, e.g. labnote -c /path/to/config.yml. If a setting is specified both in a configuration file and using a command-line switch, the option specified on the command-line will take precedence.
*Depending on how you have organized your files, this may be difficult to setup. It works best if you can normalize your directory names such that related analyses all include a similar component (e.g. ‘xx-network-analysis’).
If that is not possible or convenient, Labnote also supports manually specifying a projects categorization using hidden .labnote metafiles inside each project directory.
In addition to the automatic processing of entries that labnote normally uses to render notebook entries, directory-specific .labnote files can also be used to control the behavior and appearance of entries. These are YAML files, and should follow the format:
--- README.html: title: Custom Title pipeline.sh: title: My Interesting Analysis Pipeline
Furthermore, .labnote files can be used to specify additional entry metadata that can’t be automatically detected such as a description of the notebook entry and links to external resources such as web-pages, presentation slides, etc:
--- README.html: title: Custom Title description: Description of the notebook entry links: - http://www.google.com - research/extra/presentation.ppt
(NOTE 2016/03/02: the description and external link support haven’t been implemented yet, but should be shortly…)
The project is just getting started and is changing rapidly. Let me know if you have suggestions or would like to contribute.
The easiest way to run the unit tests for labnote is to create a virtualenv container and run the tests from within there. For example, if you have virtualenvwrapper, you can run:
git clone https://github.com/khughitt/labnote && cd labnote mkvirtualenv labnote pip install -e . pip install pytest hash -r py.test
If you already cloned the labnote repo, you can skip the first step above and simply cd to the repo directory.
The hash -r command in the above is needed after installing py.test to ensure that the virtualenv version of py.test is used, and not a system version.
To run the tests for a different version of Python, you can simply create a second virtualenv for that version of Python and repeat the process:
mkvirtualenv --python=python3.3 labnote33
Note that virtualenvwrapper is not needed to run the tests, and the commands for using the base version of virtualenv are pretty similar.
Things to be added…