Python XML processing with lxml

# - - - - -   m a i n

def main():
    """Validate one or more files against an RNC schema.

      [ if (command line arguments are valid) ->
          if (.rnc and .rng are readable, valid, and up to date) and
          (all FILE arguments are valid against that .rng) ->
            I
          else (if .rnc is readable and valid and .rng can be updated
          from the .rnc) and
          (all FILE arguments are valid against that .rng) ->
            the .rng file  :=  an RNG version of the .rnc
          else ->
            sys.stderr  +:=  error message
        else ->
            sys.stderr  +:=  error message ]
    """

Processing of the arguments is handled in Section 15.3, “rnck: checkArgs()”. We get back two items: the path to the schema, and a list of XML file names to be validated.

    #-- 1 --
    # [ if sys.argv is a valid command line ->
    #     schemaPath  :=  the SCHEMA argument
    #     fileList  :=  list of FILE arguments
    #   else ->
    #     sys.stderr  +:=  error message
    #     stop execution ]
    schemaPath, fileList = checkArgs()

Next we try to build a RelaxValidator instance from the specified schema.

    #-- 2 --
    # [ if schemaPath names a readable, valid .rng schema ->
    #       return a RelaxValidator that validates against that schema
    #   else if (schemaPath, with .rnc appended if there is no
    #   extension, names a readable, valid .rnc schema) ->
    #     if the corresponding .rng schema is readable, valid, and
    #     newer than the .rnc
    #       return a RelaxValidator that validates against the
    #       .rng schema
    #     else if (we have write access to the corresponding .rng
    #     schema) and (trang is locally installed) ->
    #       corresponding .rng schema  :=  trang's translation of
    #           the .rnc schema into .rng
    #       return a RelaxValidator that validates against
    #       translated schema
    #      else ->
    #       sys.stderr  +:=  error message
    #       stop execution ]
    validator = rnc_validate.RelaxValidator ( schemaPath )

For the logic that validates one XML file against our validator, see Section 15.7, “rnck: validateFile()”.

    #-- 3 --
    # [ sys.stderr  +:=  messages about any files from (fileList) that
    #       are unreadable or not valid against (validator) ]
    for fileName in fileList:
        validateFile ( validator, fileName )