Skip to content

Latest commit

 

History

History
139 lines (120 loc) · 7.34 KB

README.md

File metadata and controls

139 lines (120 loc) · 7.34 KB
Raw

GPS Database Processor

GPS Database Processor is designed for database validation, update and publication for the GPS (Global Pneumococcal Sequencing Project).

This tool takes the path(s) of GPS1 or GPS2 database as input:

  • -1, --gps1 - path to directory of GPS1 data
  • -2, --gps2 - path to directory of GPS2 data

It only takes one of the above in default or --check mode, and requires both in --monocle mode

It carries out several operations in the following order:

  1. Validation of columns and values of the specified directory
    • The terminal output displays any unexpected or erroneous values
    • For columns that should only contain UPPERCASE strings, any lowercase value will be converted, and the updated table will be saved in-place (unless --check mode is used)
    • If there are any critical errors, the tool will terminate its process and will not carry out the subsequent operations
  2. If run in --check mode, the operation stops here, and any case conversion will not be saved
  3. Generate table4 using inferred data based on table1, table3 and reference tables in the data directory
    • If there is a location that does not exist in data/coordinates.csv (one of the reference tables), it will stop the process
      • Except if it is running in --location mode, then it will attempt the fetch the information via Mapbox API
        • The first time this is triggered, it will ask for your Mapbox API key (access token) and save it locally at config/api_keys.confg for future use
        • For more information on the Mapbox API key (access token), please visit their documentation
  4. If not running in --monocle mode, the operation stops here
  5. Generate table_monocle.csv for Monocle
  6. Generate data.json for GPS Database Overview

 

Workflow

Workflow Diagram

 

Usage

Requirement

  • Conda
  • Git

Setup

  1. Clone this repository to your machine 
    git clone https://github.com/sanger-bentley-group/gps-database-processor.git
  2. Go into the local copy of the repository 
    cd gps-database-processor
  3. Setup Conda Environment 
    conda env create -f environment.yml

Run

  1. Pull any updates from remote repository 
    git pull
  2. Activate the Conda Environment 
    conda activate gps-db-processor
  3. Put the GPS database's three .csv source files into the directory containing the cloned repository
  4. Run the following command to validate your input files and generate the output files: 
    ./processor.py
  5. If you or the tool have updated any of the reference tables (i.e. any file in the data directory), create a PR (Pull Request) on this repository

Options

  • -1, --gps1: path to directory of GPS1 data (should contain table1.csv, table2.csv, and table3.csv of GPS1)

  • -2, --gps2: path to directory of GPS2 data (should contain table1.csv, table2.csv, and table3.csv of GPS2)

  • -c, --check: perform validation only

  • -m, --monocle: generate Monocle table and GPS Database Overview data payload from both GPS1 and GPS2

  • -l, --location: get coordinates for locations not yet exist in data/coordinates.csv via MapBox API

  • Example commands:

    ./processor.py --gps1 gps-data --check

    ./processor.py -1 gps-data -2 gps2-data -m

Reference Tables (files in the data directory)

  • alpha2_country.csv
    • Map Country in table1 to ISO 3166-1 alpha-2 code for data.json generation
    • 2 columns: alpha2, country - representing ISO 3166-1 alpha-2 code, country name respectively
    • Modify the value in the country column if it does not match the value used in Country in table1
  • coordinates.csv
    • Map the combination of Country, Region, City in table1 to Latitude, Longitude for table4 generation
    • 3 columns: Country-Region-City, Latitude, Longitude - representing Country-Region-City combination, latitude, longitude respectively
    • If a new Country-Region-City combination is found, this tool will attempt to auto-assign Latitude, Longitude to it using geopy and add them to this file
    • Modify values in the Latitude, Longitude columns if incorrect values are assigned to them by geopy
  • manifestations.csv
    • Map the combination of Clinical_manifestation and Source in table1 to Manifestation for table4 generation
    • 3 columns: Clinical_manifestation, Source, Manifestation - representing clinical manifestation, sample source, resulting manifestation respectively
    • Add new Clinical_manifestation and Source combinations and their resulting Manifestation to this file
  • non_standard_ages.csv
    • Map non-numeric values in Age_years in table1 to less_than_5years_old for table4 generation
    • 2 columns: value, less_than_5years_old - representing non-numeric age value, whether the value is less than 5 years old respectively
    • Add new non-numeric Age_years values and whether they are less_than_5years_old (Y or N) to this file
  • pcv_introduction_year.csv
    • Map Country, Year in table1 to Vaccine_period, Introduction_year, PCV_type for table4 generation
    • 3 columns: Country, PCV_type, Introduction_year - representing country name, PCV type, introduction year of the PCV respectively
    • Add new Country, PCV_type, Introduction_year to this file when that country introduces PCV or has a change of PCV type in their National Immunisation/Vaccination Programme
  • pcv_valency.csv
    • Map In_silico_serotype in table3 to all PCV types under PCV_type of this file for table4 generation
    • 2 columns: PCV_type, Serotypes - representing PCV type, list of covered serotypes by that PCV type respectively
    • Add new PCV_type, Serotypes to this file when there is a new PCV type
  • published_public_names.txt
    • Map Public_name in table1 and table3 to Published for table4 generation
    • A list of Public_name of all samples that have been published

 

Requirements & Compatibility

GPS Database requirement:

  • GPS1 v5.0+
  • GPS2 v5.0+

Tested on:

 

Credits

This project uses Open Source components. You can find the source code of their open source projects along with license information below. I acknowledge and am grateful to these developers for their contributions to open source.

pandas

geopy

draw.io / diagrams.net