bboldt commited on
Commit
036fdbb
·
1 Parent(s): 0e8a9b5

Update README.md

Browse files
Files changed (4) hide show
  1. README.md +43 -7
  2. croissant.json +1 -4
  3. util/croissant.py +3 -2
  4. util/environment.export.yml +79 -0
README.md CHANGED
@@ -1,13 +1,49 @@
1
- # Emergent Language Library
2
 
 
 
3
 
4
- ## Running code
5
 
6
- For each emergent language entry, we provide wrapper code to create
7
- a reproducible environment and run the emergent language-generating code.
8
- Environments are specified precisely in the `environment.yml` file; if you wish
9
- to edit the dependencies manually, it may be easier to start with
10
- `environment.editable.yml` instead, if it exists.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
11
 
12
  ### Git submodules
13
  This project uses git submodules to manage external dependencies. Submodules
 
1
+ # ELCC
2
 
3
+ The Emergent Language Corpus Collection is collection of corpora and metadata
4
+ from a variety of emergent communication simulations.
5
 
 
6
 
7
+ ## Using ELCC
8
+
9
+ You can clone this repository with git LFS and use the data directly or load
10
+ the data via the mlcroissant library. To install the mlcroissant library and
11
+ necessary dependencies, see the conda environment at `util/environment.yml`.
12
+ Below we show an example of loading ELCC's data via mlcroissant.
13
+ ```python
14
+ import mlcroissant as mlc
15
+
16
+ cr_url = "https://huggingface.co/datasets/bboldt/elcc/raw/main/croissant.json"
17
+ dataset = mlc.Dataset(jsonld=cr_url)
18
+
19
+ # A raw corpus of integer arrays; the corpora are named based on their paths;
20
+ # e..g., "systems/babyai-sr/data/GoToObj/corpus.json" becomes
21
+ # "babyai-sr/GoToObj".
22
+ records = dataset.records(record_set="babyai-sr/GoToObj")
23
+ # System-level metadata
24
+ records = dataset.records(record_set="system-metadata")
25
+ # Raw JSON string for system metadata; some fields aren't handled well by
26
+ # Croissant, so you can access them here if need be.
27
+ records = dataset.records(record_set="system-metadata-raw")
28
+ # Corpus metadata, specifically metrics generated by ELCC's analyses
29
+ records = dataset.records(record_set="corpus-metadata")
30
+ # Raw corpus metadata
31
+ records = dataset.records(record_set="corpus-metadata-raw")
32
+
33
+ # `records` can now be iterated through to access the individual elements.
34
+ ```
35
+
36
+
37
+ ## Developing
38
+
39
+ ### Running individual EC systems
40
+ For each emergent language entry, we provide wrapper code (in
41
+ `systems/*/code/`) to create a reproducible environment and run the emergent
42
+ language-generating code. Environments are specified precisely in the
43
+ `environment.yml` file; if you wish to edit the dependencies manually, it may
44
+ be easier to start with `environment.editable.yml` instead, if it exists.
45
+ Next, either run or look at `run.sh` or `run.py` to see the commands necessary
46
+ to produce to the corpora.
47
 
48
  ### Git submodules
49
  This project uses git submodules to manage external dependencies. Submodules
croissant.json CHANGED
@@ -49,7 +49,7 @@
49
  "description": "ELCC is a collection of emergent language corpora with accompanying metadata and analyses.",
50
  "conformsTo": "http://mlcommons.org/croissant/1.0",
51
  "citeAs": "@misc{boldt2024elcc,\n title = \"{ELCC}: the Emergent Language Corpus Collection\",\n author = \"Brendon Boldt and David Mortensen\",\n year = 2024,\n url = \"https://huggingface.co/datasets/bboldt/elcc\",\n}",
52
- "datePublished": "2024-06-13 07:19:21.844134+00:00",
53
  "keywords": [
54
  "emergent communication",
55
  "emergent language",
@@ -1142,9 +1142,6 @@
1142
  },
1143
  "extract": {
1144
  "fileProperty": "fullpath"
1145
- },
1146
- "transform": {
1147
- "replace": "a"
1148
  }
1149
  }
1150
  },
 
49
  "description": "ELCC is a collection of emergent language corpora with accompanying metadata and analyses.",
50
  "conformsTo": "http://mlcommons.org/croissant/1.0",
51
  "citeAs": "@misc{boldt2024elcc,\n title = \"{ELCC}: the Emergent Language Corpus Collection\",\n author = \"Brendon Boldt and David Mortensen\",\n year = 2024,\n url = \"https://huggingface.co/datasets/bboldt/elcc\",\n}",
52
+ "datePublished": "2024-06-13 07:25:44.022899+00:00",
53
  "keywords": [
54
  "emergent communication",
55
  "emergent language",
 
1142
  },
1143
  "extract": {
1144
  "fileProperty": "fullpath"
 
 
 
1145
  }
1146
  }
1147
  },
util/croissant.py CHANGED
@@ -310,7 +310,8 @@ def test() -> None:
310
  # records = dataset.records(record_set="babyai-sr/GoToObj")
311
  # records = dataset.records(record_set="system-metadata")
312
  # records = dataset.records(record_set="system-metadata-raw")
313
- records = dataset.records(record_set="corpus-metadata")
 
314
 
315
  for i, x in enumerate(records):
316
  print(i, x)
@@ -321,4 +322,4 @@ def test() -> None:
321
 
322
  if __name__ == "__main__":
323
  save_metadata()
324
- test()
 
310
  # records = dataset.records(record_set="babyai-sr/GoToObj")
311
  # records = dataset.records(record_set="system-metadata")
312
  # records = dataset.records(record_set="system-metadata-raw")
313
+ # records = dataset.records(record_set="corpus-metadata")
314
+ records = dataset.records(record_set="corpus-metadata-raw")
315
 
316
  for i, x in enumerate(records):
317
  print(i, x)
 
322
 
323
  if __name__ == "__main__":
324
  save_metadata()
325
+ # test()
util/environment.export.yml ADDED
@@ -0,0 +1,79 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ name: elcc
2
+ channels:
3
+ - conda-forge
4
+ dependencies:
5
+ - _libgcc_mutex=0.1=conda_forge
6
+ - _openmp_mutex=4.5=2_gnu
7
+ - annotated-types=0.7.0=pyhd8ed1ab_0
8
+ - black=24.4.2=py312h7900ff3_0
9
+ - bzip2=1.0.8=hd590300_5
10
+ - c-ares=1.28.1=hd590300_0
11
+ - ca-certificates=2024.6.2=hbcca054_0
12
+ - click=8.1.7=unix_pyh707e725_0
13
+ - colorama=0.4.6=pyhd8ed1ab_0
14
+ - git=2.45.1=pl5321hef9f9f3_0
15
+ - git-lfs=3.5.1=ha770c72_0
16
+ - jinja2=3.1.4=pyhd8ed1ab_0
17
+ - joblib=1.4.2=pyhd8ed1ab_0
18
+ - keyutils=1.6.1=h166bdaf_0
19
+ - krb5=1.21.2=h659d440_0
20
+ - ld_impl_linux-64=2.40=hf3520f5_2
21
+ - libblas=3.9.0=22_linux64_openblas
22
+ - libcblas=3.9.0=22_linux64_openblas
23
+ - libcurl=8.8.0=hca28451_0
24
+ - libedit=3.1.20191231=he28a2e2_2
25
+ - libev=4.33=hd590300_2
26
+ - libexpat=2.6.2=h59595ed_0
27
+ - libffi=3.4.2=h7f98852_5
28
+ - libgcc-ng=13.2.0=h77fa898_7
29
+ - libgfortran-ng=13.2.0=h69a702a_7
30
+ - libgfortran5=13.2.0=hca663fb_7
31
+ - libgomp=13.2.0=h77fa898_7
32
+ - libiconv=1.17=hd590300_2
33
+ - liblapack=3.9.0=22_linux64_openblas
34
+ - libnghttp2=1.58.0=h47da74e_1
35
+ - libnsl=2.0.1=hd590300_0
36
+ - libopenblas=0.3.27=pthreads_h413a1c8_0
37
+ - libsqlite=3.45.3=h2797004_0
38
+ - libssh2=1.11.0=h0841786_0
39
+ - libstdcxx-ng=13.2.0=hc0a3c3a_7
40
+ - libuuid=2.38.1=h0b41bf4_0
41
+ - libxcrypt=4.4.36=hd590300_1
42
+ - libzlib=1.3.1=h4ab18f5_1
43
+ - markupsafe=2.1.5=py312h98912ed_0
44
+ - mypy=1.10.0=py312h9a8786e_0
45
+ - mypy_extensions=1.0.0=pyha770c72_0
46
+ - ncurses=6.5=h59595ed_0
47
+ - numpy=1.26.4=py312heda63a1_0
48
+ - openssl=3.3.1=h4ab18f5_0
49
+ - packaging=24.0=pyhd8ed1ab_0
50
+ - pandas=2.2.2=py312h1d6d2e6_1
51
+ - pandas-stubs=2.2.2.240603=pyhd8ed1ab_0
52
+ - pathspec=0.12.1=pyhd8ed1ab_0
53
+ - pcre2=10.43=hcad00b1_0
54
+ - perl=5.32.1=7_hd590300_perl5
55
+ - pip=24.0=pyhd8ed1ab_0
56
+ - platformdirs=4.2.2=pyhd8ed1ab_0
57
+ - psutil=5.9.8=py312h98912ed_0
58
+ - pydantic=2.7.3=pyhd8ed1ab_0
59
+ - pydantic-core=2.18.4=py312h4413252_0
60
+ - python=3.12.3=hab00c5b_0_cpython
61
+ - python-dateutil=2.9.0=pyhd8ed1ab_0
62
+ - python-tzdata=2024.1=pyhd8ed1ab_0
63
+ - python_abi=3.12=4_cp312
64
+ - pytz=2024.1=pyhd8ed1ab_0
65
+ - pyyaml=6.0.1=py312h98912ed_1
66
+ - readline=8.2=h8228510_1
67
+ - setuptools=70.0.0=pyhd8ed1ab_0
68
+ - six=1.16.0=pyh6c4a22f_0
69
+ - tk=8.6.13=noxft_h4845f30_101
70
+ - tqdm=4.66.4=pyhd8ed1ab_0
71
+ - types-pytz=2024.1.0.20240417=pyhd8ed1ab_0
72
+ - typing-extensions=4.12.1=hd8ed1ab_0
73
+ - typing_extensions=4.12.1=pyha770c72_0
74
+ - tzdata=2024a=h0c530f3_0
75
+ - wheel=0.43.0=pyhd8ed1ab_1
76
+ - xz=5.2.6=h166bdaf_0
77
+ - yaml=0.2.5=h7f98852_2
78
+ - zstd=1.5.6=ha6fb4c9_0
79
+