v3.0.1

Fix

• pipeline.py: Correct filters of jaccard (5679713)

v3.0.0

Fix

• pipeline.py: Filter jaccard calculations (ba0f97b)
• run.py: Correct passing of control (04d5c5a)
• pipeline.py: Correct grouped Jaccard calculation (c2c100c)
• pipeline.py: Correct enfc implementation (529fa49)

Breaking

• Previously, the pipeline only calculated the jaccard scores for a single, hard-coded pairing. This has been corrected. All users should update to continue to get the correct results. (c2c100c)
• Prior implementations had calculated the ENFC for class by taking the mean and st.dev of the individual lipid ENFC for all lipids in a class. While not incorrect, this was inconsistent with prior work of this framework. Closes #23. (529fa49)

Documentation

• README.md: Update output desciption (366dd98)
• data_handling.py: Correct docstrings (bfadda5)
• README.md: Add version number disclaimer (b2b78dd)

v2.0.2

Fix

• jaccard.py: Catch zerodivision error (9278005)

Documentation

• pyproject.toml: Update classifiers (f43ccab)

v2.0.1

Fix

• pipeline.py: Exclude control condition (88313ed)

v2.0.0

Feature

• examples: Provide minimum working examples (c0575f8)
• cli: Calculate enfc against control (f9d7e6b)

Fix

• parser.py: Update default parameters (878df07)

Breaking

• Current users should update any scripts and configs to use the new --compartment flag rather than the existing --tissue flag. (5ca9aed)
• Users should update their configuration files to ensure that they specify 0.2 as the threshold if they wish to continue using the old value. (878df07)
• Previously, if multiple fold change calculations were desired, the user had to call lta multiple times. Now, a single call produces all fold change calculations for a set of experimental conditions. This feature could have been implemented by having the user specify all pairs they were interested in. While this could introduce more flexibility (ie it would allow for multiple "controls"), it also introduces more overhead for both the user and maintainer. Additionally, for the vast majority of experimental designs, there will be only one control. Closes #17. (f9d7e6b)

Documentation

• README.md: Enumerate defaults (3404845)
• pipeline.py: Update docstrings for new functions (35e98d0)
• LICENSE: Update copyright year (d4fd7a0)
• README.md: Update usage instructions (cdc2f09)
• README.md: Add python installation instructions (4fc9b83)

v1.0.0

Feature

• cli: Change phenotype to group (6b91841)
• cli: Generalise data reading (d5ce9f5)

Breaking

• Change "phenotype" to "group". Existing users will need to make sure that any scripts and config files use this new flag. (6b91841)
• This introduces 2 new command line flags and changes how the data is processed. Past scripts are not guaranteed to run correctly - unless, by chance, they used the defaults - so a breaking release is required. (d5ce9f5)

Documentation

• README.md: Update usage for new flags (75772f7)

v0.12.5

Fix

• lta.py: Remove colons from filenames (3e4388c)

v0.12.4

Fix

• pipeline.py: Respect metadata locations (e906e87)

Documentation

• repo: Use Traffic throughout (af3c4d1)
• README.md: Update badges (57718d4)

v0.12.3

Fix

• pyproject.toml: Correct version number (0b01aea)
• pyproject.toml: Correct included packages (d603a6c)

v0.12.2

v0.12.2