DHParser 0.9.0

DHParser - Parser generator and DSL-construction-kit

DHParser

DHParser - A parser generator and domain specific language (DSL) construction kit for the Digit Humanities

Features

• Fail-tolerant parsing

• Customizable error reporting

• Unit testing and debugging of grammars

• Language-server-support (https://microsoft.github.io/language-server-protocol/)

• Beginner friendly

• Digital Humanities oriented (optional XML-workflows!), but open for any application domain

• Python-based

Ease of use

key_value_store.py:

# A mini-DSL for a key value store
from DHParser import *

# specify the grammar of your DSL in EBNF-notation
grammar = '''@ drop = whitespace, token
key_store   = ~ { entry }             
entry       = key "=" value 
key         = /\w+/~                  # Scannerless parsing: Use regular
value       = /\"[^"\n]*\"/~          # expressions wherever you like'''

# generating a parser is almost as simple as compiling a regular expression    
parser_factory = grammar_provider(grammar)
parser = parser_factory()             # parser factory for thread-safety

Now, parse some text and extract the data from the Python-shell:

>>> from key_value_store import parser
>>> text = '''
        title    = "Odysee 2001"
        director = "Stanley Kubrick"
    '''         
>>> data = parser(text)  
>>> for entry in data.select('entry'):
        print(entry['key'], entry['value'])

title "Odysee 2001"
director "Stanley Kubrick"

Or, serialize as XML:

>>> print(data.as_xml())

<key_store>
  <entry>
    <key>title</key>
    <value>"Odysee 2001"</value>
  </entry>
  <entry>
    <key>director</key>
    <value>"Stanley Kubrick"</value>
  </entry>
</key_store>    

License

DHParser is open source software under the Apache 2.0 License.

Copyright 2016-2019 Eckhart Arnold, Bavarian Academy of Sciences and Humanities

Licensed under the Apache License, Version 2.0 (the "License"); you may not use this file except in compliance with the License. You may obtain a copy of the License at

https://www.apache.org/licenses/LICENSE-2.0e

Unless required by applicable law or agreed to in writing, software distributed under the License is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the License for the specific language governing permissions and limitations under the License.

Author

Author: Eckhart Arnold, Bavarian Academy of Sciences Email: [e-mail]

Installation

You can install DHParser from the Python package index (https://pypi.org)*:

python -m pip install --user --upgrade DHParser

However, as the software is still in an early beta stage, it is recommended that you pull it directly from gitlab (see below).

Post-Installaton

It is recommended that you install the regex-module (https://bitbucket.org/mrabarnett/mrab-regex). If present, DHParser will use regex instead of the built-in re-module for regular expressions. regex is faster and more powerful than re.

In order to speed up DHParser even more, it is recommended that you run the buildpackages.sh-script (or buildpackages.bat on Windows-systems) after installation. This script compiles the stringview.py-module to C-Code with the cython Python to C compiler (https://cython.org). This requires that you have cython and a C-compiler installed on your system.

Sources

Find the sources on gitlab.lrz.de/badw-it/DHParser . Get them with:

git clone https://gitlab.lrz.de/badw-it/DHParser

There exists a mirror of this repository on github: https://github.com/jecki/DHParser Be aware, though, that the github-mirror may occasionally lag behind a few commits.

Please contact me, if you are intested in contributing to the development or just using DHParser.

Disclaimer

DHParser is still in development stage. While it is already quite mature, there may still be changes to the API.

Purpose

DHParser is a parser-combinator-based parsing and compiling infrastructure for domain specific languages (DSL) in Digital Humanities projects. It leverages the power of Domain specific languages for the Digital Humanities.

Domain specific languages are widespread in computer sciences, but seem to be underused in the Digital Humanities. While DSLs are sometimes introduced to Digital-Humanities-projects as practical adhoc-solution, these solutions are often somewhat "quick and dirty". In other words they are more of a hack than a technology. The purpose of DHParser is to introduce DSLs as a technology to the Digital Humanities. It is based on the well known technology of EBNF-based parser generators, but employs the more modern form called "parsing expression grammar" and parser combinators as a variant of the classical recursive descent parser.

Why another parser generator? There are plenty of good parser generators out there, e.g. Añez's grako parser generator, Eclipse XText. However, DHParser is intended as a tool that is specifically geared towards digital humanities applications, while most existing parser generators come from compiler construction toolkits for programming languages. While I expect DSLs in computer science and DSLs in the Digital Humanities to be quite similar as far as the technological realization is concerned, the use cases, requirements and challenges are somewhat different. For example, in the humanities annotating text is a central use case, which is mostly absent in computer science treatments. These differences might sooner or later require to develop the DSL-construction toolkits in a different direction. Also DHParser emphasizes and evolutionary development model for grammars with unit-testing support, which fits the typical use cases in DH where DSLs evolve in a discussion process between technicians and humanists. Because the users of DSLs in the humanities are not necessarily very technically mindes people, DHParser supports the construction of fail-tolerant parsers with good error reporting in terms of locating the errors at the right spot and giving useful error messages.

Also, DHParser shall (in the future) serve as a teaching tool, which influences some of its design decisions such as, for example, clearly separating the parsing, syntax-tree-transformation and compilation stages. Finally, DHParser is intended as a tool to experiment with. One possible research area is, how non context-free grammars such as the grammars of TeX or CommonMark can be described with declarative langauges in the spirit of but beyond EBNF, and what extensions of the parsing technology are necessary to capture such languages.

Primary use case at the Bavarian Academy of Sciences and Humanities (for the time being): A DSL for the "Mittellateinische Wörterbuch"!

Further (intended) use cases are:

• LaTeX -> XML/HTML conversion. See this discussion on why an EBNF-parser for the complete TeX/LaTeX-grammar is not possible.
• CommonMark and other DSLs for cross media publishing of scientific literature, e.g. journal articles. (Common Mark and Markdown also go beyond what is feasible with pure EBNF-based-parsers.)
• EBNF itself. DHParser is already self-hosting ;-)
• XML-parser, just for the fun of it ;-)
• Digital and cross-media editions
• Digital dictionaries

For a simple self-test run dhparser.py from the command line. This compiles the EBNF-Grammer in examples/EBNF/EBNF.ebnf and outputs the Python-based parser class representing that grammar. The concrete and abstract syntax tree as well as a full and abbreviated log of the parsing process will be stored in a sub-directory named "LOG".

Introduction

See Introduction.md for the motivation and an overview how DHParser works or jump right into the Step by Step Guide to learn how to setup and use DHParser.

References and Acknowledment

Juancarlo Añez: grako, a PEG parser generator in Python, 2017. URL: bitbucket.org/apalala/grako

Eckhart Arnold: Domänenspezifische Notationen. Eine (noch) unterschätzte Technologie in den Digitalen Geisteswissenschaften, Präsentation auf dem dhmuc-Workshop: Digitale Editionen und Auszeichnungssprachen, München 2016. Short-URL: tiny.badw.de/2JVT

Brian Ford: Parsing Expression Grammars: A Recognition-Based Syntactic Foundation, Cambridge Massachusetts, 2004. Short-URL:t1p.de/jihs

Richard A. Frost, Rahmatullah Hafiz and Paul Callaghan: Parser Combinators for Ambiguous Left-Recursive Grammars, in: P. Hudak and D.S. Warren (Eds.): PADL 2008, LNCS 4902, pp. 167–181, Springer-Verlag Berlin Heidelberg 2008.

Elizabeth Scott and Adrian Johnstone, GLL Parsing, in: Electronic Notes in Theoretical Computer Science 253 (2010) 177–189, dotat.at/tmp/gll.pdf

Dominikus Herzberg: Objekt-orientierte Parser-Kombinatoren in Python, Blog-Post, September, 18th 2008 on denkspuren. gedanken, ideen, anregungen und links rund um informatik-themen, short-URL: t1p.de/bm3k

Dominikus Herzberg: Eine einfache Grammatik für LaTeX, Blog-Post, September, 18th 2008 on denkspuren. gedanken, ideen, anregungen und links rund um informatik-themen, short-URL: t1p.de/7jzh

Dominikus Herzberg: Uniform Syntax, Blog-Post, February, 27th 2007 on denkspuren. gedanken, ideen, anregungen und links rund um informatik-themen, short-URL: t1p.de/s0zk

John MacFarlane, David Greenspan, Vicent Marti, Neil Williams, Benjamin Dumke-von der Ehe, Jeff Atwood: CommonMark. A strongly defined, highly compatible specification of Markdown, 2017. commonmark.org

Stefan Müller: DSLs in den digitalen Geisteswissenschaften, Präsentation auf dem dhmuc-Workshop: Digitale Editionen und Auszeichnungssprachen, München 2016. Short-URL: tiny.badw.de/2JVy

Markus Voelter, Sbastian Benz, Christian Dietrich, Birgit Engelmann, Mats Helander, Lennart Kats, Eelco Visser, Guido Wachsmuth: DSL Engineering. Designing, Implementing and Using Domain-Specific Languages, 2013. dslbook.org/

Christopher Seaton: A Programming Language Where the Syntax and Semantics are Mutuable at Runtime, University of Bristol 2007, chrisseaton.com/katahdin/katahdin.pdf

Vegard Øye: General Parser Combinators in Racket, 2012, epsil.github.io/gll/