Skip to content

Job request: 18286

Organisation:
The London School of Hygiene & Tropical Medicine
Workspace:
openprompt-hrqol
ID:
ic46oipx2shk66mr
View repo View project.yaml

This page shows the technical details of what happened when the authorised researcher Alasdair Henderson requested one or more actions to be run against real patient data within a secure environment.

By cross-referencing the list of jobs with the pipeline section below, you can infer what security level the outputs were written to.

The output security levels are:

  • highly_sensitive
    • Researchers can never directly view these outputs
    • Researchers can only request code is run against them
  • moderately_sensitive
    • Can be viewed by an approved researcher by logging into a highly secure environment
    • These are the only outputs that can be requested for public release via a controlled output review service.

Jobs

Pipeline

Show project.yaml 
version: '3.0'

expectations:
 population_size: 10000

actions:

  create_dummy_data: 
    run: >
      ehrql:v0
        create-dummy-tables 
        analysis/model_questions/process_baseline.py output/dummydata 
        -- 
        --day=0
    outputs: 
      highly_sensitive:
        openprompt_dummy: output/dummydata/open_prompt.csv

  edit_dummy_data:
    run: > 
      r:latest
        analysis/dummy_data_editing/edit_automatic_dummy_data.R
    needs: [create_dummy_data]
    outputs: 
      highly_sensitive: 
        openprompt_dummy_edited: output/dummydata/dummy_edited/open_prompt.csv

  generate_openprompt_baseline: 
    run: >
      databuilder:v0
        generate-dataset 
        analysis/model_questions/process_baseline.py 
        --output output/openprompt_baseline.csv
        --dummy-tables output/dummydata/dummy_edited  
        --
        --day=0
    needs: [edit_dummy_data]
    outputs:
      highly_sensitive:
        openprompt_baseline: output/openprompt_baseline.csv

  generate_openprompt_survey1: 
    run: >
      databuilder:v0
        generate-dataset 
        analysis/model_questions/process_research.py 
        --output output/openprompt_survey1.csv
        --dummy-tables output/dummydata/dummy_edited
        --
        --day=0
    needs: [edit_dummy_data]
    outputs:
      highly_sensitive:
        openprompt_survey1: output/openprompt_survey1.csv

  generate_openprompt_survey2: 
    run: >
      databuilder:v0
        generate-dataset 
        analysis/model_questions/process_research.py 
        --output output/openprompt_survey2.csv
        --dummy-tables output/dummydata/dummy_edited
        --
        --day=30
    needs: [edit_dummy_data]
    outputs:
      highly_sensitive:
        openprompt_survey2: output/openprompt_survey2.csv

  generate_openprompt_survey3: 
    run: >
      databuilder:v0
        generate-dataset 
        analysis/model_questions/process_research.py 
        --output output/openprompt_survey3.csv
        --dummy-tables output/dummydata/dummy_edited
        --
        --day=60
    needs: [edit_dummy_data]
    outputs:
      highly_sensitive:
        openprompt_survey3: output/openprompt_survey3.csv

  generate_openprompt_survey4: 
    run: >
      databuilder:v0
        generate-dataset 
        analysis/model_questions/process_research.py 
        --output output/openprompt_survey4.csv
        --dummy-tables output/dummydata/dummy_edited
        --
        --day=90
    needs: [edit_dummy_data]
    outputs:
      highly_sensitive:
        openprompt_survey4: output/openprompt_survey4.csv

  combine_openprompt:
    run: >
      r:latest analysis/001_datacombine.R
    needs: [generate_openprompt_baseline, generate_openprompt_survey1, generate_openprompt_survey2, generate_openprompt_survey3, generate_openprompt_survey4]
    outputs: 
      highly_sensitive: 
        openprompt_combined: output/openprompt_raw.csv.gz
      moderately_sensitive:
        openprompt_raw_skim: output/data_properties/op_raw_skim.txt
        openprompt_raw_tab: output/data_properties/op_raw_tabulate.txt
        openprompt_mapped_skim: output/data_properties/op_mapped_skim.txt
        openprompt_mapped_tab: output/data_properties/op_mapped_tabulate.txt
        check_days_after_baseline: output/data_properties/sample_day_lags.pdf
        indexdates: output/data_properties/index_dates.pdf
        table1: output/tab1_baseline_description.html
        raw_summ_base_s: output/data_properties/op_baseline_skim.txt
        raw_summ_base_t: output/data_properties/op_baseline_tabulate.txt
        raw_summ_survey1_s: output/data_properties/op_survey1_skim.txt
        raw_summ_survey1_t: output/data_properties/op_survey1_tabulate.txt
        raw_summ_survey2_s: output/data_properties/op_survey2_skim.txt
        raw_summ_survey2_t: output/data_properties/op_survey2_tabulate.txt
        raw_summ_survey3_s: output/data_properties/op_survey3_skim.txt
        raw_summ_survey3_t: output/data_properties/op_survey3_tabulate.txt
        raw_summ_survey4_s: output/data_properties/op_survey4_skim.txt
        raw_summ_survey4_t: output/data_properties/op_survey4_tabulate.txt

  # generate_openprompt_plus_tpp: 
  #   run: >
  #     databuilder:v0
  #       generate-dataset analysis/dataset_definition_openprompt.py --output output/openprompt_raw_plus_tpp.csv.gz
  #   needs: [create_dummy_openprompt_data]
  #   outputs:
  #     highly_sensitive:
  #       openprompt_tpp_combined: output/openprompt_raw_plus_tpp.csv.gz

  # quick_summ_data:
  #   run: >
  #     r:latest
  #       analysis/010_table1.R
  #   needs: [generate_openprompt_plus_tpp]
  #   outputs:
  #     highly_sensitive:
  #       cleandata: output/cleaned_data.gz.parquet
  #     moderately_sensitive:
  #       table1: output/tab1_baseline_description.html
  #       longcovid_dates: output/longcovid_dates.pdf

Timeline

  • Created:

  • Started:

  • Finished:

  • Runtime: 00:03:01

These timestamps are generated and stored using the UTC timezone on the TPP backend.

Job request

Status
Succeeded
Backend
TPP
Workspace
openprompt-hrqol
Requested by
Alasdair Henderson
Branch
main
Force run dependencies
Yes
Git commit hash
0f46c49
Requested actions
  • run_all

Code comparison

Compare the code used in this job request