Skip to content

Validating with pyx12

pyx12 checks a file against the implementation guide. oxedi can run it for you and return its findings as ordinary diagnostics.

pip install "oxedi[pyx12]"

parse never calls pyx12. validate does, when you ask.

Validate a file

This example breaks the segment count of the example file, so both oxedi and pyx12 have something to report:

from pathlib import Path

import oxedi
from oxedi.pyx12 import validate

data = Path("remittance.835").read_bytes().replace(b"SE*58*", b"SE*57*")
result = oxedi.parse(data)
findings = result.diagnostics + validate(data)    # bytes, a path or a binary file object

for d in sorted(findings, key=lambda d: d.level):
    print(d.origin, d.code, d)                    # "oxedi" or "pyx12"; pyx12's own code
    if d.segment is not None:
        start, end = result.document[d.segment].span  # the segment's bytes in your file

oxedi None SNIP 1 · SE01 declares "57" but the count is 58 · segment #59, element 1 · at interchange#1/group#1/transaction#1 · datum "57"
pyx12 4 SNIP 1 · SE count of 57 for SE02=0001 is wrong. I count 58 (reported by pyx12, code 4) · segment #59, element 1 · at interchange#1/group#1/transaction#1 · datum "57"
pyx12 10 SNIP 2 · Data element "Implementation Convention Reference" (ST03) is marked as Not Used (reported by pyx12, code 10) · segment #2, element 3 · at interchange#1/group#1/transaction#1 · datum "005010X221A1"

pyx12 also reports ST03, which its map marks as not used. It does so on the unedited file too.

  • validate returns oxedi.Diagnostic, the type parse returns. The two lists mix, sort by level and filter by origin.
  • validate returns its findings in segment order.
  • A file with no ISA raises oxedi.ParseError, as parse does.

What a pyx12 finding holds

validate reports every error pyx12's engine records, read from the error tree pyx12 builds:

  • Interchange, group and transaction errors. These are level 1.
  • Segment and element errors, those of the envelope segments included. These are level 2.

A pyx12 finding has:

  • kind == "External".
  • origin == "pyx12".
  • code set to pyx12's error code.
  • path naming the loops open at its segment, as parse names them.

Where a finding points

  • A count or control number that does not match lands on the trailer segment, at the element holding it, with that value as datum.
  • A duplicate or missing control structure lands on the header, at the control number when that is the offending value.
  • When pyx12 gives no value, datum is the element's bytes, or the segment id for a finding with no element.

When pyx12 cannot read a file

A file pyx12 cannot read, or an exception inside pyx12, comes back as one level 1 finding instead of a traceback. It has no code, and its rule starts with could not finish validating.

  • A file pyx12 rejects points at its ISA segment, with the ISA bytes as datum.
  • A failure after the last segment names the last segment pyx12 completed.
  • A failure while reading names the segment pyx12 was processing.
Details: logging

pyx12 also logs what it finds, under the pyx12 logger. On first use, validate attaches a logging.NullHandler to that logger when it has no handler. So the records are not printed by default, and they reach any handler you configure. This also applies to code that runs pyx12 directly in the same process. No level, propagation or logging.disable setting is changed.