Skip to content
This repository

HTTPS clone URL

Subversion checkout URL

You can clone with HTTPS or Subversion.

Download ZIP
tree: b64edcb01d

Fetching latest commit…

Cannot retrieve the latest commit at this time

README.md

Template Parser

Parse ASCII files by example

Basic usage

First, create a template. In this example I'm parsing an ASCII purchase order of some sort.

template = TemplateParser.compile_template([ "?<:>REPORT NUMBER: ? ", " INVENTORY NO: #number REF NUMBER: #refer_number LOGICAL UNIT: #unit_number ? ", " CATALOG CODE: :cat_code :sub_code VENDOR NO: :vendor ", " TYPE CODE: :lpr FLAG: ? BILL FLAG: <:bill_flag>#_ SHIPPING FROM: ? SHIPPING TO: ? ", " FROM-LOCATION: :address_code TO LOCATION: :to_address_code ", " :address_name ? ", " :address_1 ? ", " :address_2 ? ", " :address_city :address_postal ? ", " USER NAME: :user_name SERIAL NUMBER: :serial " ].join("\n"))

Let's break down the template a little bit.

The first thing to note is that it's not hard to imagine what the date we're parsing will actually look like. That's because the field definitions go in the physical locations where the data should be, while the rest of the report appears unchanged. While parsing a report, if a single character is out of place, parsing will fail with a detailed error message. I've found that by failing fast I've been able to find the edge cases easily and get perfect parsing results on literally gigabytes of generated reports.

There are a few different types of fields visible here as well. Some start with #, indicating a numeric field. Most start with : and look like symbols, indicating a text field. There are also some ?'s indicating that something may appear there but we'll ignore it. Finally there are some zero-width field names which look like <:name>:_ which can appear where the field name would otherwise not fit. A variation of that is where <:> can be used by itself as a 0-width delimiter to prevent a field from being too long.

Side note: typically I would use ruby here doc string notation but here am concatinating an array of strings to help demonstrate that whitespace is significant. That way I can just copy in an example of the report I'm interested in and carve out my field definitions directly.

An array of line matchers are returned from the compile_template method.

Using the array of line matchers (template)

Does the given line have a match in any of the lines in the template?

TemplateParser.match_template?(template, line)

Get the results of matching any line in the template to the given line

TemplateParser.match_template(template, line, file_position_metadata) { |matcher, converted_data, raw_data| }

Process all given lines against the template in order

TemplateParser.process_lines(template, lines, file_position_metadata)

Return true if process_lines would run successfully on the given lines for the given template.

TemplateParser.lines_match_template?(template, lines)

Using individual line matchers

Does the given line match the given line matcher?

TemplateParser.match_line?(line_matcher, line)

Process a given line against a given line matcher

TemplateParser.process_line(line_matcher, line, file_position_metadata) { |matcher, converted_data, raw_data| }

Something went wrong with that request. Please try again.