fix(parsers): honour HL7 v2 MSH delimiters, repetitions and escape sequences

[kevincostner17]

Summary

The HL7 v2 ER7 parser hard-coded | and ^, so it ignored the delimiters a message declares in MSH-1/MSH-2. It never split ~ repetitions and never decoded escape sequences. PID-3 12345~98765 became the patient id, \S\/\T\ stayed in values, and a #-delimited message parsed to 0 messages.

• Each MSH sets its message's delimiters: field separator from seg[3], component/repetition/escape/subcomponent from MSH-2 (missing ones default to ^~\&). They apply until the next MSH.
• Segment ids are read with seg[:3], and a known segment must be followed by the current field separator. Unknown segments are still reported by name.
• Single-valued columns (ids, names, codes, dates, units, status) take the first repetition.
• OBX-5 (observation value) is a repeating field, so it keeps every repetition, joined with ~. A custom repetition character is re-emitted as ~.
• Fields are split before escapes are decoded, so an escaped delimiter such as \R\ never splits a value. \F\ \S\ \T\ \R\ \E\ are decoded using the message's delimiters. Other escapes (\H\, \N\, \Xhh\, \.br\) are kept verbatim.
• Whole-field values and message_type use the standard ^ / & / ~ separators whatever the message declared.
• Standard |^~\& messages without escapes, and without repetitions outside OBX-5, give the same frames, metadata and warnings as before.

Most of the hl7v2.py line count is ruff format on a file that was not formatted on main.

Tests

New tests/parsers/test_hl7v2_delimiters.py:

• the issue repro (first repetition, escapes, # field separator)
• custom MSH#$*@! delimiters, and a partial MSH-2 that falls back to defaults
• first repetition for single-valued columns
• OBX-5 keeping every repetition, with standard delimiters and with a custom repetition character
• \R\ decoded after the split rather than split on
• each escape, plus unknown and unterminated escapes, and escapes using the message's own delimiters
• escaped delimiters not splitting components
• three messages with alternating delimiters in one input
• a PIDX prefix treated as unknown
• a standard multi-message fixture compared against frames from the previous parser

Verification

• ruff check . passes; changed files are ruff format-clean
• mypy src/freshdata: no issues in 202 source files
• pytest -m "not online and not large": py3.12 4483 passed / 6 skipped; py3.9 4479 passed / 10 skipped

Closes #261

[coderabbitai]

Important

• 🔍 Trigger review

This repository does not receive automatic reviews because it has fewer than 10 stars.

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Advanced

Run ID: 2294033a-a40a-4e93-bc0e-c8cfa0fb083f

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[github-actions]

FreshData benchmark report — performance

• freshdata: ?
• python: ?
• platform: ?

fixture	n_rows	n_cols	p50 s	p95 s	peak MB	repair %	false-repair %	preserve %	trust	monotonic	export %

Authored-code reduction (Metric 6)