README.md 12.3 KB
Newer Older
1 2
DHParser
========
Eckhart Arnold's avatar
Eckhart Arnold committed
3

4 5
DHParser - A parser generator and domain specific language (DSL) construction 
           kit for the Digit Humanities
eckhart's avatar
eckhart committed
6 7 8 9 10


Features
--------

eckhart's avatar
eckhart committed
11 12 13 14 15 16 17 18 19 20 21
* *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
eckhart's avatar
eckhart committed
22

eckhart's avatar
eckhart committed
23
* *Python-based*
Eckhart Arnold's avatar
Eckhart Arnold committed
24 25


26 27 28 29 30 31 32 33 34 35 36 37
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 
eckhart's avatar
eckhart committed
38
    key         = /\w+/~                  # Scannerless parsing: Use regular
39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74
    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>    


Eckhart Arnold's avatar
Eckhart Arnold committed
75 76 77
License
-------

78
DHParser is open source software under the [Apache 2.0 License](https://www.apache.org/licenses/LICENSE-2.0).
79

80
Copyright 2016-2019  Eckhart Arnold, Bavarian Academy of Sciences and Humanities
Eckhart Arnold's avatar
Eckhart Arnold committed
81

82 83 84
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
85

86
    https://www.apache.org/licenses/LICENSE-2.0e
Eckhart Arnold's avatar
Eckhart Arnold committed
87

88 89 90 91 92
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
93

eckhart's avatar
eckhart committed
94 95 96 97 98 99 100

Author
------

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

101

102 103 104 105 106 107 108 109 110 111
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).

eckhart's avatar
eckhart committed
112

113 114 115 116 117 118 119 120 121 122 123
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
124
`stringview.py`-module to C-Code with the cython Python to C compiler
125 126 127 128
(https://cython.org). This requires that you have cython and a
C-compiler installed on your system.


129 130 131
Sources
-------

132
Find the sources on [gitlab.lrz.de/badw-it/DHParser](https://gitlab.lrz.de/badw-it/DHParser) .
133
Get them with:
134

135
    git clone https://gitlab.lrz.de/badw-it/DHParser
136

137 138 139 140
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
141
Please contact me, if you are intested in contributing to the
142
development or just using DHParser.
Eckhart Arnold's avatar
Eckhart Arnold committed
143

eckhart's avatar
eckhart committed
144

Eckhart Arnold's avatar
Eckhart Arnold committed
145 146 147
Disclaimer
----------

eckhart's avatar
eckhart committed
148 149 150
DHParser is still in development stage. While it is already quite mature, 
there may still be changes to the API.

Eckhart Arnold's avatar
Eckhart Arnold committed
151

152 153 154
Purpose
-------

eckhart's avatar
eckhart committed
155 156 157 158
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.
159

160
Domain specific languages are widespread in
161 162
computer sciences, but seem to be underused in the Digital Humanities.
While DSLs are sometimes introduced to Digital-Humanities-projects as
163 164 165 166 167 168 169 170 171 172 173
[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
174
generators out there, e.g. [Añez's grako parser generator][Añez_2017],
175
[Eclipse XText][XText_Website]. However, DHParser is
176 177
intended as a tool that is specifically geared towards digital
humanities applications, while most existing parser generators come
178 179
from compiler construction toolkits for programming languages.
While I expect DSLs in computer science and DSLs in the Digital
180
Humanities to be quite similar as far as the technological realization
181 182 183 184
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
eckhart's avatar
eckhart committed
185 186 187 188 189 190 191 192 193 194
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,
195
DHParser shall (in the future) serve as a teaching tool, which
196 197
influences some of its design decisions such as, for example, clearly
separating the parsing, syntax-tree-transformation and compilation
198
stages. Finally, DHParser is intended as a tool to experiment with.  One
199 200 201 202 203 204 205 206 207 208
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/)"!
209 210 211

Further (intended) use cases are:

212 213 214 215 216 217 218
* 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.)
219
* EBNF itself. DHParser is already self-hosting ;-)
220
* XML-parser, just for the fun of it ;-)
221
* Digital and cross-media editions
222
* Digital dictionaries
Eckhart Arnold's avatar
Eckhart Arnold committed
223

224 225 226 227
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
228 229
parsing process will be stored in a sub-directory named "LOG".

eckhart's avatar
eckhart committed
230

231 232 233
Introduction
------------

234 235 236 237
See [Introduction.md](https://gitlab.lrz.de/badw-it/DHParser/blob/master/Introduction.md) for the 
motivation and an overview how DHParser works or jump right into the 
[Step by Step Guide](https://gitlab.lrz.de/badw-it/DHParser/blob/master/documentation_src/StepByStepGuide.rst) to
learn how to setup and use DHParser.
Eckhart Arnold's avatar
Eckhart Arnold committed
238

eckhart's avatar
eckhart committed
239

240
References and Acknowledment
241 242
----------

243 244
Juancarlo Añez: grako, a PEG parser generator in Python, 2017. URL:
[bitbucket.org/apalala/grako][Añez_2017]
245

di68kap's avatar
di68kap committed
246
[Añez_2017]: https://bitbucket.org/apalala/grako
247 248


249 250 251 252
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
253
München 2016. Short-URL: [tiny.badw.de/2JVT][Arnold_2016]
254

di68kap's avatar
di68kap committed
255
[Arnold_2016]: https://f.hypotheses.org/wp-content/blogs.dir/1856/files/2016/12/EA_Pr%C3%A4sentation_Auszeichnungssprachen.pdf
256

257 258
Brian Ford: Parsing Expression Grammars: A Recognition-Based Syntactic
Foundation, Cambridge
259
Massachusetts, 2004. Short-URL:[t1p.de/jihs][Ford_2004]
260 261 262

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

263
[Ford_20XX]: http://bford.info/packrat/
264 265 266 267 268 269

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.

270 271 272 273 274 275
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

276 277
Dominikus Herzberg: Objekt-orientierte Parser-Kombinatoren in Python,
Blog-Post, September, 18th 2008 on denkspuren. gedanken, ideen,
278
anregungen und links rund um informatik-themen, short-URL:
279
[t1p.de/bm3k][Herzberg_2008a]
280 281 282 283 284

[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
285
links rund um informatik-themen, short-URL:
286
[t1p.de/7jzh][Herzberg_2008b]
287 288 289 290 291

[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
292
informatik-themen, short-URL: [t1p.de/s0zk][Herzberg_2007]
293 294 295 296 297

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

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

298 299 300 301
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]
302 303 304

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

305 306 307 308
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]
309 310 311

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

312 313
Markus Voelter, Sbastian Benz, Christian Dietrich, Birgit Engelmann,
Mats Helander, Lennart Kats, Eelco Visser, Guido Wachsmuth:
314
DSL Engineering. Designing, Implementing and Using Domain-Specific Languages, 2013.
315 316 317 318 319 320 321 322 323 324 325 326
[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
327 328

[voelter_2013]: http://dslbook.org/
329 330

[tex_stackexchange_no_bnf]: http://tex.stackexchange.com/questions/4201/is-there-a-bnf-grammar-of-the-tex-language
331 332

[tex_stackexchange_latex_parsers]: http://tex.stackexchange.com/questions/4223/what-parsers-for-latex-mathematics-exist-outside-of-the-tex-engines
333 334

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