The Week 4
As per the schedule, we met for the fourth time on 4th June. Again, it was a wonderful meeting. I and Andreas arrived a few minutes before, so we discussed the project's progress first, I told him we had completed the Milestone 1.3, and then we looked at a few issues which I had solved. We also discussed about the COBRApy community, which was a little less responsive towards our project. Matthias arrived soon. He suggested we should ask on the Gitter channel of the COBRApy that where should we put our introductory mail. I posted it there and am waiting for the reply.
The mentors then started reviewing the code. The code was overall good, they both appreciated it. Matthias suggested a few changes, or more generally, a few "python tricks", to make the code look more better. I later made the changes and pushed them on Github. We further went to look after one of the most important parts of the project, handling of annotation inside COBRApy, which is right now very minimally supported. This is the one part that we think would not be backward compatible because the current way of handling the annotation suffers from many issues. So even if we try to extend the current format, we would only be able to handle the issue of losing the biology qualifiers, but would never be able to handle more complex annotations. The current version only specifies that these annotations had to be of dictionary type and don't put any further restriction. But as discussed in the last blog, we are going to have separate classes for handling this annotation part.
I had already started working on this annotation class last week because I was supposedly required to start at least some working on it. I had almost completely implemented the History class and the KeyValuePair class. I initially thought that I should put them as inner classes inside the main Annotation class, but the mentors further suggested me to make a separate package for these meta information classes and put all of them into it. So I made the package inside the COBRApy's core directory. The mentors then suggested to name the classes in a way so that they sound more verbose and don't just have simple names (like the Annotation class should be named as MetaData). So I would be naming them accordingly. I further showed them the format that I have in my mind for this annotation class, here's the discussion which went on for this format. Both of them almost agreed, just provided a few suggestions. We decided to put the resource URI as complete, instead of breaking it into parts, because many other URI/URL's may not conform to the pattern of breaking up. We decided to have a method that can split the URI, if possible, into parts and return the value so that users can use it if they want. This would help to have a lossless conversion of data.
So we finally decided to complete this meta-data part in the following week. That marked the ending of the meeting.
Thank you for the read. I hope you are enjoying it.
See you next week.
The mentors then started reviewing the code. The code was overall good, they both appreciated it. Matthias suggested a few changes, or more generally, a few "python tricks", to make the code look more better. I later made the changes and pushed them on Github. We further went to look after one of the most important parts of the project, handling of annotation inside COBRApy, which is right now very minimally supported. This is the one part that we think would not be backward compatible because the current way of handling the annotation suffers from many issues. So even if we try to extend the current format, we would only be able to handle the issue of losing the biology qualifiers, but would never be able to handle more complex annotations. The current version only specifies that these annotations had to be of dictionary type and don't put any further restriction. But as discussed in the last blog, we are going to have separate classes for handling this annotation part.
I had already started working on this annotation class last week because I was supposedly required to start at least some working on it. I had almost completely implemented the History class and the KeyValuePair class. I initially thought that I should put them as inner classes inside the main Annotation class, but the mentors further suggested me to make a separate package for these meta information classes and put all of them into it. So I made the package inside the COBRApy's core directory. The mentors then suggested to name the classes in a way so that they sound more verbose and don't just have simple names (like the Annotation class should be named as MetaData). So I would be naming them accordingly. I further showed them the format that I have in my mind for this annotation class, here's the discussion which went on for this format. Both of them almost agreed, just provided a few suggestions. We decided to put the resource URI as complete, instead of breaking it into parts, because many other URI/URL's may not conform to the pattern of breaking up. We decided to have a method that can split the URI, if possible, into parts and return the value so that users can use it if they want. This would help to have a lossless conversion of data.
So we finally decided to complete this meta-data part in the following week. That marked the ending of the meeting.
Thank you for the read. I hope you are enjoying it.
See you next week.
Comments
Post a Comment