README.md 10.3 KB
Newer Older
1 2
DHParser
========
Eckhart Arnold's avatar
Eckhart Arnold committed
3

4
A parser-combinator-based parsing and compiling infrastructure for domain
5
specific languages (DSL) in Digital Humanities projects.
Eckhart Arnold's avatar
Eckhart Arnold committed
6 7 8 9 10 11 12

Author: Eckhart Arnold, Bavarian Academy of Sciences
Email:  arnold@badw.de

License
-------

13
DHParser is open source software under the [Apache 2.0 License](https://www.apache.org/licenses/LICENSE-2.0)
14

Eckhart Arnold's avatar
Eckhart Arnold committed
15 16
Copyright 2016-2017  Eckhart Arnold, Bavarian Academy of Sciences and Humanities

17 18 19
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
Eckhart Arnold's avatar
Eckhart Arnold committed
20

21
    https://www.apache.org/licenses/LICENSE-2.0
Eckhart Arnold's avatar
Eckhart Arnold committed
22

23 24 25 26 27
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.
Eckhart Arnold's avatar
Eckhart Arnold committed
28

29
**Exception**: The module ``DHParser/foreign_typing.py`` was directly taken from the
30
Python 3.5 source code in order for DHParser to be backwards compatible
31
with Python 3.4. The module ``DHParser/foreign_typing.py`` is licensed under the
32 33
[Python Software Foundation License Version 2](https://docs.python.org/3.5/license.html)

34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54
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
di68kap's avatar
di68kap committed
55
`stringview.py`-module to C-Code with the cython Python to C compiler
56 57 58 59
(https://cython.org). This requires that you have cython and a
C-compiler installed on your system.


60 61 62
Sources
-------

63
Find the sources on [gitlab.lrz.de/badw-it/DHParser](https://gitlab.lrz.de/badw-it/DHParser) .
64
Get them with:
65

66
    git clone https://gitlab.lrz.de/badw-it/DHParser
67

68 69 70 71
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.

Eckhart Arnold's avatar
Eckhart Arnold committed
72
Please contact me, if you are intested in contributing to the
73
development or just using DHParser.
Eckhart Arnold's avatar
Eckhart Arnold committed
74 75 76 77 78 79 80 81 82 83 84

Disclaimer
----------

DHParser is still in an early development stage. While it is definitaly
usable, features may be dropped or added without notice and class or
function names changed in future versions. The API is NOT YET STABLE!

Use it for testing an evaluation, but not in an production environment
or contact me first, if you intend to do so.

85 86 87
Purpose
-------

88
DHParser leverages the power of Domain specific languages for the
89 90
Digital Humanities.

91
Domain specific languages are widespread in
92 93
computer sciences, but seem to be underused in the Digital Humanities.
While DSLs are sometimes introduced to Digital-Humanities-projects as
94 95 96 97 98 99 100 101 102 103 104
[practical adhoc-solution][Müller_2016], 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][Arnold_2016] to the Digital Humanities. It is
based on the well known technology of [EBNF][ISO_IEC_14977]-based
parser generators, but employs the more modern form called
"[parsing expression grammar][Ford_2004]" and
[parser combinators][Ford_20XX] as a variant of the classical
recursive descent parser.

Why another parser generator? There are plenty of good parser
105
generators out there, e.g. [Añez's grako parser generator][Añez_2017],
106
[Eclipse XText][XText_Website]. However, DHParser is
107 108
intended as a tool that is specifically geared towards digital
humanities applications, while most existing parser generators come
109 110
from compiler construction toolkits for programming languages.
While I expect DSLs in computer science and DSLs in the Digital
111
Humanities to be quite similar as far as the technological realization
112 113 114 115
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
116 117
DSL-construction toolkits in a different direction. Also,
DHParser shall (in the future) serve as a teaching tool, which
118 119
influences some of its design decisions such as, for example, clearly
separating the parsing, syntax-tree-transformation and compilation
120
stages. Finally, DHParser is intended as a tool to experiment with.  One
121 122 123 124 125 126 127 128 129 130
possible research area is, how non
[context-free grammars](https://en.wikipedia.org/wiki/Context-free_grammar)
such as the grammars of [TeX][tex_stackexchange_no_bnf] or
[CommonMark][MacFarlane_et_al_2017] 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](http://www.mlw.badw.de/)"!
131 132 133

Further (intended) use cases are:

134 135 136 137 138 139 140
* LaTeX -> XML/HTML conversion. See this
  [discussion on why an EBNF-parser for the complete TeX/LaTeX-grammar][tex_stackexchange_no_bnf]
  is not possible.
* [CommonMark][MacFarlane_et_al_2017] 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.)
141
* EBNF itself. DHParser is already self-hosting ;-)
142
* XML-parser, just for the fun of it ;-)
143
* Digital and cross-media editions
144
* Digital dictionaries
Eckhart Arnold's avatar
Eckhart Arnold committed
145

146 147 148 149
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
150 151 152 153 154
parsing process will be stored in a sub-directory named "LOG".

Introduction
------------

Eckhart Arnold's avatar
Eckhart Arnold committed
155
see [Introduction.md](https://gitlab.lrz.de/badw-it/DHParser/blob/master/Introduction.md)
Eckhart Arnold's avatar
Eckhart Arnold committed
156

157
References and Acknowledment
158 159
----------

160 161
Juancarlo Añez: grako, a PEG parser generator in Python, 2017. URL:
[bitbucket.org/apalala/grako][Añez_2017]
162

di68kap's avatar
di68kap committed
163
[Añez_2017]: https://bitbucket.org/apalala/grako
164 165


166 167 168 169
Eckhart Arnold: Domänenspezifische Notationen. Eine (noch)
unterschätzte Technologie in den Digitalen Geisteswissenschaften,
Präsentation auf dem
[dhmuc-Workshop: Digitale Editionen und Auszeichnungssprachen](https://dhmuc.hypotheses.org/workshop-digitale-editionen-und-auszeichnungssprachen),
di68kap's avatar
di68kap committed
170
München 2016. Short-URL: [tiny.badw.de/2JVT][Arnold_2016]
171

di68kap's avatar
di68kap committed
172
[Arnold_2016]: https://f.hypotheses.org/wp-content/blogs.dir/1856/files/2016/12/EA_Pr%C3%A4sentation_Auszeichnungssprachen.pdf
173

174 175
Brian Ford: Parsing Expression Grammars: A Recognition-Based Syntactic
Foundation, Cambridge
176
Massachusetts, 2004. Short-URL:[t1p.de/jihs][Ford_2004]
177 178 179

[Ford_2004]: https://pdos.csail.mit.edu/~baford/packrat/popl04/peg-popl04.pdf

180
[Ford_20XX]: http://bford.info/packrat/
181 182 183 184 185 186

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.

187 188 189 190 191 192
Elizabeth Scott and Adrian Johnstone, GLL Parsing,
in: Electronic Notes in Theoretical Computer Science 253 (2010) 177–189,
[dotat.at/tmp/gll.pdf][scott_johnstone_2010]

[scott_johnstone_2010]: http://dotat.at/tmp/gll.pdf

193 194
Dominikus Herzberg: Objekt-orientierte Parser-Kombinatoren in Python,
Blog-Post, September, 18th 2008 on denkspuren. gedanken, ideen,
195
anregungen und links rund um informatik-themen, short-URL:
196
[t1p.de/bm3k][Herzberg_2008a]
197 198 199 200 201

[Herzberg_2008a]: http://denkspuren.blogspot.de/2008/09/objekt-orientierte-parser-kombinatoren.html

Dominikus Herzberg: Eine einfache Grammatik für LaTeX, Blog-Post,
September, 18th 2008 on denkspuren. gedanken, ideen, anregungen und
202
links rund um informatik-themen, short-URL:
203
[t1p.de/7jzh][Herzberg_2008b]
204 205 206 207 208

[Herzberg_2008b]: http://denkspuren.blogspot.de/2008/09/eine-einfache-grammatik-fr-latex.html

Dominikus Herzberg: Uniform Syntax, Blog-Post, February, 27th 2007 on
denkspuren. gedanken, ideen, anregungen und links rund um
209
informatik-themen, short-URL: [t1p.de/s0zk][Herzberg_2007]
210 211 212 213 214

[Herzberg_2007]: http://denkspuren.blogspot.de/2007/02/uniform-syntax.html

[ISO_IEC_14977]: http://www.cl.cam.ac.uk/~mgk25/iso-14977.pdf

215 216 217 218
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][MacFarlane_et_al_2017]
219 220 221

[MacFarlane_et_al_2017]: http://commonmark.org/

222 223 224 225
Stefan Müller: DSLs in den digitalen Geisteswissenschaften,
Präsentation auf dem
[dhmuc-Workshop: Digitale Editionen und Auszeichnungssprachen](https://dhmuc.hypotheses.org/workshop-digitale-editionen-und-auszeichnungssprachen),
München 2016. Short-URL: [tiny.badw.de/2JVy][Müller_2016]
226 227 228

[Müller_2016]: https://f.hypotheses.org/wp-content/blogs.dir/1856/files/2016/12/Mueller_Anzeichnung_10_Vortrag_M%C3%BCnchen.pdf

229 230
Markus Voelter, Sbastian Benz, Christian Dietrich, Birgit Engelmann,
Mats Helander, Lennart Kats, Eelco Visser, Guido Wachsmuth:
231
DSL Engineering. Designing, Implementing and Using Domain-Specific Languages, 2013.
232 233 234 235 236 237 238 239 240 241 242 243
[dslbook.org/][Voelter_2013]

Christopher Seaton: A Programming Language Where the Syntax and Semantics
are Mutuable at Runtime, University of Bristol 2007,
[chrisseaton.com/katahdin/katahdin.pdf][seaton_2007]

Vegard Øye: General Parser Combinators in Racket, 2012,
[epsil.github.io/gll/][vegard_2012]

[vegard_2012]: https://epsil.github.io/gll/ 

[seaton_2007]: http://chrisseaton.com/katahdin/katahdin.pdf
244 245

[voelter_2013]: http://dslbook.org/
246 247

[tex_stackexchange_no_bnf]: http://tex.stackexchange.com/questions/4201/is-there-a-bnf-grammar-of-the-tex-language
248 249

[tex_stackexchange_latex_parsers]: http://tex.stackexchange.com/questions/4223/what-parsers-for-latex-mathematics-exist-outside-of-the-tex-engines
250 251

[XText_website]: https://www.eclipse.org/Xtext/