Skip to contents

Export Validation Flags to MongoDB

export_validation_flags.Rd

Exports validation flags to MongoDB, keeping the decisions reviewers made in the Peskas Management Platform or in KoboToolbox (read beforehand with coasts::review_decisions()).

Usage

export_validation_flags(
  conf = NULL,
  asset_id = c("v1", "v2", "v3"),
  all_flags = NULL,
  validation_statuses = NULL
)

Arguments

conf

Configuration object from read_config() containing MongoDB connection parameters and survey-specific settings

asset_id

Character string specifying which WF survey version to process. Must be one of "v1", "v2" or "v3". Determines which configuration to use from conf$ingestion$wf_{asset_id}. Default is "v1".

all_flags

Data frame containing all validation flags with columns: submission_id, submitted_by, submission_date, alert_flag

validation_statuses

Reviewers' decisions from coasts::review_decisions(), with columns submission_id, validation_status, validated_at, validated_by

Value

Invisible NULL. The function pushes data to MongoDB as a side effect.

Details

The function performs the following steps:

  1. Joins validation flags with KoboToolbox validation statuses

  2. Identifies manual human approvals (excluding system username)

  3. Preserves manual human decisions while updating system-generated statuses

  4. Creates both wide and long format datasets for different reporting needs

  5. Pushes results directly to MongoDB collections

Validation Status Logic:

  • If submission has flags AND validated_by is system username: set to "not_approved"

  • If submission has no flags AND validated_by is system username: set to "approved"

  • If validated_by is NOT system username: preserve existing status (a reviewer's approval or rejection)

Note

This function is called internally by validate_wf_surveys() and should not typically be called directly. It requires:

  • Valid configuration with MongoDB connection string

  • Survey-specific configuration under conf$ingestion$wf_{asset_id}

  • System username configured to identify automated vs. manual validations

MongoDB Collections

The function pushes to two MongoDB collections, named with the survey's KoBo asset id:

surveys_flags-asset_id

Wide format with one row per submission including validation status and flags

enumerators_stats-asset_id

Long format with one row per flag per submission for enumerator statistics

Examples

if (FALSE) { # \dontrun{
# Called internally by validate_wf_surveys()
export_validation_flags(
  conf = conf,
  asset_id = "v1",
  all_flags = flags_combined,
  validation_statuses = validation_statuses
)
} # }