From 23ad25fee1d1d3c7fefd1a87e06d301913457e34 Mon Sep 17 00:00:00 2001 From: Tim Kucera Date: Fri, 24 Nov 2023 11:31:46 +0100 Subject: [PATCH] file structure --- proteinshake/backend/{structure.py => protein.py} | 0 proteinshake/readme.md | 15 +++++++++++++++ 2 files changed, 15 insertions(+) rename proteinshake/backend/{structure.py => protein.py} (100%) create mode 100644 proteinshake/readme.md diff --git a/proteinshake/backend/structure.py b/proteinshake/backend/protein.py similarity index 100% rename from proteinshake/backend/structure.py rename to proteinshake/backend/protein.py diff --git a/proteinshake/readme.md b/proteinshake/readme.md new file mode 100644 index 00000000..5c4e1c75 --- /dev/null +++ b/proteinshake/readme.md @@ -0,0 +1,15 @@ +Filestructure: + +**backend:** Collects and processes raw pdb files from various databases. Creates Collections. + **adapters:** Database adapters to PDB, AFDB, etc. + **protein:** The protein object specification. + **collection:** A set of protein objects. + **database:*** Unifying database mirror that allows query access to create a collection. +**frontend:** High level specifications for datasets and tasks. + **datasets:** Load a collection and apply transforms. + **tasks:** Various ML prediction problems. Takes a dataset and rearranges the data (sklearn-style X_train, y_train, ...) such that it can be fed to the model (X) and evaluated (y). + **evaluators:** Groups relevant metrics for a given class of problems (e.g. classification or regression). + **splitters:** Takes a dataset and generates split indices, either from the data/labels itself or from other resources. +**datasets:** Some relevant collections of proteins, for preprocessed hosting. Each dataset covers some biological topic. +**tasks:** The actually implemented biological tasks for the end-user, consisting of a frontend.task in combination with a specific dataset. +**transforms:** Various functions to transform proteins, representations, and frameworks. \ No newline at end of file