normalizer

Normalizer

A data integration layer receives records from multiple sources with inconsistent formatting: US dates (MM/DD/YYYY) vs. ISO dates, phone numbers with or without country codes, addresses with abbreviations. The normalizer pipeline applies source-specific rules to produce uniformly formatted output records.

Pipeline

[nrm_detect_format]
     |
     v
     <SWITCH>
       |-- json -> [nrm_convert_json]
       |-- xml -> [nrm_convert_xml]
       |-- csv -> [nrm_convert_csv]
       +-- default -> [nrm_convert_json]
     |
     v
[nrm_output]

Workflow inputs: rawInput, sourceSystem

Workers

NrmConvertCsvWorker (task: nrm_convert_csv)

• Writes canonical

NrmConvertJsonWorker (task: nrm_convert_json)

• Writes canonical

NrmConvertXmlWorker (task: nrm_convert_xml)

• Writes canonical

NrmDetectFormatWorker (task: nrm_detect_format)

• Trims whitespace
• Reads rawInput. Writes detectedFormat

NrmOutputWorker (task: nrm_output)

• Writes normalized, originalFormat, outputFormat

---

20 tests | Workflow: nrm_normalizer | Timeout: 60s

See RUNNING.md for setup and usage.