BiocBuildDB is an R package that provides tools for managing and querying a database of Bioconductor build reports. These reports contain information about the build status of Bioconductor packages across different platforms and R versions.
Data flow
Each day, the Bioconductor build reports are processed to generate structured tables of build results. These tables are stacked to generate a longitudinal record of build results available as parquet files. The data flow for the entire process is outlined in the diagram below.
flowchart TD
%% =========================
%% Phase A: Daily extraction
%% =========================
A[Bioconductor<br/>Daily Build Reports]
-->|daily| B[Parse & Normalize<br/>Build Metadata]
B -->|md5 hashed<br/>by build report| C1[info.csv]
B -->|md5 hashed<br/>by build report| C2[build_summary.csv]
B -->|md5 hashed<br/>by build report| C3[propagation_status.csv]
%% =========================
%% Phase B: Longitudinal stacking
%% =========================
C1 -->|append new report| D1[info.parquet<br/>partitioned by date]
C2 -->|append new report| D2[build_summary.parquet<br/>partitioned by date]
C3 -->|append new report| D3[propagation_status.parquet<br/>partitioned by date]
%% =========================
%% Storage layer
%% =========================
D1 --> E[(Object Storage\nS3 / GCS)]
D2 --> E
D3 --> E
%% =========================
%% Storage layer
%% =========================
E -->|Usage| U[BiocBuildDB R Package]
E -->|Usage| DASH[Third party dashboards]
E -->|Usage| ANALYSIS[Analytics]
%% =========================
%% Styling
%% =========================
classDef source fill:#eef,stroke:#446;
classDef daily fill:#efe,stroke:#484;
classDef parquet fill:#ffe,stroke:#aa4;
classDef storage fill:#fdf6e3,stroke:#b58900;
class A source;
class C1,C2,C3 daily;
class D1,D2,D3 parquet;
class E storage;Installation
You can install the development version of BiocBuildDB from GitHub with:
# install.packages("BiocManager")
BiocManager("seandavi/BiocBuildDB")