Skip to content

Latest commit

 

History

History
174 lines (144 loc) · 7.74 KB

CHANGELOG.md

File metadata and controls

174 lines (144 loc) · 7.74 KB
Raw

Change Log

All notable changes to this project will be documented in this file.

2.0 2017-04-09

Added

  • Added tool to generate mock DDI data, for both badness and agonism
  • Added exported data from running the mock DDI data generation tool

Fixed

  • Mock DDI data is deterministic across different runs of the app
  • Display correct spacing time units in frontend

1.3 2017-04-02

Added

  • Added UI sections for displaying the different types of DDI sequencing categories identified by the system: parallel, sequential, repeated alternative, and alternative non-DDI
  • Added the ability to specify a delay in PML files

Changed

  • Redesigned the UI to use tabs for the different logical sections of the application: Analysis, Warnings, and Errors
  • The syntax error UI now includes contextual line highlighting

1.2 2017-03-26

Added

  • Added a features document for Release 2.
  • Parser now reports unnamed PML constructs. These are then displayed to the user in a similar manner to other warnings.
  • Parser now reports PML construct name clashes. These are then displayed to the user in a similar manner to other warnings.

Changed

  • Updated how warnings are displayed. Where possible, warnings now contain a preview of the section of the PML file that caused the warning. This is currently implemented for Unnamed Constructs and PML Construct Name Clashes. We intend to use a similar design for other errors and warnings in the future.

Fixed

  • Running automated integration tests will now correctly test the release matching the current tag that is checked out.

1.1 2017-03-19

Added

  • Added support for PML TX file downloads. Valid pml files uploaded to the system for transformation can now be downloaded via the UI.

1.0 2017-03-12

Changed

  • Clarified installation instructions after feedback from client. Client left feedback after release 0.3 indicating that our current README was unhelpful. We have greatly simplified it and moved more non-essential information into separate documents.
  • Updated user interface with loading icon to provide visual representation of the request's duration.
  • Changed the homepage to update when a file is selected not when it is analysed, based on client feedback.
  • Updated the representation of drugs in PML after client feedback. Our initial representation meant that users had to look up the CHEBI/DINTO identifier for a drug. Now users can specify drugs with a new drug { "name" } construct.
  • Added UI section to display drugs that were specified in the PML but that are not included in DINTO
  • Updated user interface to enable live updates in which the results of the PML parsing and DINTO lookup are displayed as soon as they are found before the DDIs have been calculated
  • Updated how analysis results are displayed to make them more readable. Previously we would display the JSON responses from the API to the user. Now we take the JSON responses and create a readable message which is then displayed to the user.
  • Updated the testing instructions. The client indicated that he would like clear instructions for manual acceptance tests. We have updated the testing instructions in FEATURES.md accordingly. We also removed references to Panacea's automated unit tests. A section was added about running our new end-to-end tests instead, as these are a much more thorough test of the system.

Added

  • install-docker.sh for easy docker installation on Ubuntu 16.04. Part of simplifying the installation steps.
  • Created Athloi: This service runs our end to end Cucumber tests using Selenium WebDriver and Firefox. This service also runs during continuous integration to ensure there are no regressions to the end user interface between commits/releases.

Fixed

  • Fixed a bug where analysis requests would time out.
  • Fixed bug where results of the previous DDI analysis would not be cleared in the event there was an error with the current analysis

0.3 2017-03-05

Added

  • Created Chiron: a new service for Querying DINTO built with Apache Jena. DINTO is turned into a triple store at compile time, so boot time is greatly improved. Similarly, query times are much improved by the triple store
  • DDI analysis is now presented to users. After validating a PML file, Panacea queries Asclepius to find any DDIs between drugs in the file. This is then displayed to the user in the UI.

Removed

  • DINTO from Asclepius. Chiron houses DINTO data now. Asclepius now proxies queries to Chiron.
  • Removed code that waited for DINTO to be loaded. Previously, Asclepius would take several minutes to load DINTO into memory so Panacea would have to poll it to see if it was ready. Chiron can load DINTO from its triple store instantly, so the waiting is no longer required.

0.2 2017-02-26

Added

  • docker-compose. As we now have more than one docker service docker-compose is used to easily coordinate them. docker-compose can be installed as per the installation instructions in the README
  • Panacea: a new docker service that is responsible for validating PML and serving the UI. The UI is a web app
  • Parser for PML. Added a custom lexer and parser for PML. Panacea will expose a HTTP API for uploading files to be parsed.
  • Continuous Integration testing. Automated testing has been added. CI runs whenever a commit is pushed. Merging into master and our iteration branches requires CI to pass.
  • Asclepius: a new service for querying DINTO. This service loads DINTO into memory when it boots and exposes a HTTP API for making queries. Panacea will use this API to provide users with DDI feedback.
  • PML upload. Added a form that allows users to upload a file. The file is sent to Panacea for analysis.
  • PML error & warning highlights. The results of file analysis are now reported back to the user.
  • Asclepius DDI endpoint. Asclepius now exposes an endpoint that takes a list of drugs and returns any DDIs between those drugs.

Removed

  • Removed Pathways. The old docker service that contained peos has been removed. It is replaced by Asclepius and Panacea.
  • Removed submodules. Previously Pathways used submodules as part of its docker build process. Now Asclepius clones DINTO and checks out a specific revision instead.

0.1 2017-02-12

Fixed

  • Fixed installation instructions. Added a workaround for a DNS issue on the TCD network.

Added

  • Instructions on how to run docker without sudo on Ubuntu.

0.0 2017-02-05

Added

  • Peos submodule. For now we'll use Peos to interact with PML. This will probably be replaced by a custom parser later as we really only need the parsing.
  • DINTO submodule. The project depends on DINTO to provide information about drug-drug interactions (DDIs)
  • Dockerfile for building project. At the moment, the project consists of one docker service: pathways.
  • Installation instructions in README