AndrewRWilliams commited on
Commit
2370f7d
·
verified ·
1 Parent(s): 59884a4

Upload folder using huggingface_hub

Browse files
Files changed (3) hide show
  1. README.md +0 -32
  2. configs.json +38 -0
  3. time_mmd_multi.py +10 -13
README.md CHANGED
@@ -1,34 +1,2 @@
1
- ---
2
- pretty_name: Time-MMD (multi-config)
3
- tags:
4
- - time-series
5
- - forecasting
6
- - multivariate
7
- - multimodal
8
- task_categories:
9
- - time-series-forecasting
10
- language:
11
- - en
12
- license: mit
13
- annotations_creators:
14
- - no-annotation
15
- source_datasets:
16
- - original
17
- multilinguality: monolingual
18
- size_categories:
19
- - 1K<n<10K
20
- ---
21
-
22
- # Time-MMD Multi-Config
23
-
24
- Each subfolder under `data/` is a config (e.g., `Economy_predLen_24`, `Health_predLen_48`, ...).
25
- Load one with:
26
-
27
- ```python
28
- from datasets import load_dataset
29
- ds = load_dataset("your-username/time-mmd-multiconfig", name="<config_name>")
30
- print(ds) # DatasetDict with train/validation/test
31
-
32
-
33
  # Time-MMD Multi-Config
34
  Each subfolder in `data/` is a config.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
  # Time-MMD Multi-Config
2
  Each subfolder in `data/` is a config.
configs.json ADDED
@@ -0,0 +1,38 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ "newyork_aqi_day_predLen_192",
3
+ "newyork_aqi_day_predLen_336",
4
+ "newyork_aqi_day_predLen_48",
5
+ "newyork_aqi_day_predLen_96",
6
+ "unadj_unemploymentrate_all_processed_predLen_10",
7
+ "unadj_unemploymentrate_all_processed_predLen_12",
8
+ "unadj_unemploymentrate_all_processed_predLen_6",
9
+ "unadj_unemploymentrate_all_processed_predLen_8",
10
+ "us_femagrant_month_predLen_10",
11
+ "us_femagrant_month_predLen_12",
12
+ "us_femagrant_month_predLen_6",
13
+ "us_femagrant_month_predLen_8",
14
+ "us_fluratio_week_predLen_12",
15
+ "us_fluratio_week_predLen_24",
16
+ "us_fluratio_week_predLen_36",
17
+ "us_fluratio_week_predLen_48",
18
+ "us_gasolineprice_week_predLen_12",
19
+ "us_gasolineprice_week_predLen_24",
20
+ "us_gasolineprice_week_predLen_36",
21
+ "us_gasolineprice_week_predLen_48",
22
+ "us_precipitation_month_predLen_10",
23
+ "us_precipitation_month_predLen_12",
24
+ "us_precipitation_month_predLen_6",
25
+ "us_precipitation_month_predLen_8",
26
+ "us_retailbroilercomposite_month_predLen_10",
27
+ "us_retailbroilercomposite_month_predLen_12",
28
+ "us_retailbroilercomposite_month_predLen_6",
29
+ "us_retailbroilercomposite_month_predLen_8",
30
+ "us_tradebalance_month_predLen_10",
31
+ "us_tradebalance_month_predLen_12",
32
+ "us_tradebalance_month_predLen_6",
33
+ "us_tradebalance_month_predLen_8",
34
+ "us_vmt_month_predLen_10",
35
+ "us_vmt_month_predLen_12",
36
+ "us_vmt_month_predLen_6",
37
+ "us_vmt_month_predLen_8"
38
+ ]
time_mmd_multi.py CHANGED
@@ -7,14 +7,11 @@ _DESCRIPTION = (
7
  _CITATION = ""
8
 
9
 
10
- def _list_configs():
11
- root = os.path.dirname(__file__)
12
- data_root = os.path.join(root, "data")
13
- if not os.path.isdir(data_root):
14
- return []
15
- return sorted(
16
- d for d in os.listdir(data_root) if os.path.isdir(os.path.join(data_root, d))
17
- )
18
 
19
 
20
  class _Cfg(datasets.BuilderConfig):
@@ -25,7 +22,7 @@ class _Cfg(datasets.BuilderConfig):
25
 
26
  class TimeMMDMulti(datasets.GeneratorBasedBuilder):
27
  BUILDER_CONFIGS = [
28
- _Cfg(name=n, description=f"Subset: {n}") for n in _list_configs()
29
  ]
30
  DEFAULT_CONFIG_NAME = BUILDER_CONFIGS[0].name if BUILDER_CONFIGS else None
31
 
@@ -51,19 +48,19 @@ class TimeMMDMulti(datasets.GeneratorBasedBuilder):
51
  def _split_generators(self, dl_manager):
52
  root = os.path.dirname(__file__)
53
  d = os.path.join(root, "data", self.config.subdir)
54
- files = {
55
  "train": os.path.join(d, "train.jsonl"),
56
  "validation": os.path.join(d, "validation.jsonl"),
57
  "test": os.path.join(d, "test.jsonl"),
58
  }
59
  gens = []
60
- for name, path in files.items():
61
- if os.path.exists(path):
62
  gens.append(
63
  datasets.SplitGenerator(
64
  name=getattr(datasets.Split, name.upper()),
65
  gen_kwargs={
66
- "filepath": path,
67
  "split_name": name,
68
  "domain": self.config.subdir,
69
  },
 
7
  _CITATION = ""
8
 
9
 
10
+ def _config_names():
11
+ here = os.path.dirname(__file__)
12
+ cfg_path = os.path.join(here, "configs.json")
13
+ with open(cfg_path, "r", encoding="utf-8") as f:
14
+ return json.load(f)
 
 
 
15
 
16
 
17
  class _Cfg(datasets.BuilderConfig):
 
22
 
23
  class TimeMMDMulti(datasets.GeneratorBasedBuilder):
24
  BUILDER_CONFIGS = [
25
+ _Cfg(name=n, description=f"Subset: {n}") for n in _config_names()
26
  ]
27
  DEFAULT_CONFIG_NAME = BUILDER_CONFIGS[0].name if BUILDER_CONFIGS else None
28
 
 
48
  def _split_generators(self, dl_manager):
49
  root = os.path.dirname(__file__)
50
  d = os.path.join(root, "data", self.config.subdir)
51
+ paths = {
52
  "train": os.path.join(d, "train.jsonl"),
53
  "validation": os.path.join(d, "validation.jsonl"),
54
  "test": os.path.join(d, "test.jsonl"),
55
  }
56
  gens = []
57
+ for name, p in paths.items():
58
+ if os.path.exists(p):
59
  gens.append(
60
  datasets.SplitGenerator(
61
  name=getattr(datasets.Split, name.upper()),
62
  gen_kwargs={
63
+ "filepath": p,
64
  "split_name": name,
65
  "domain": self.config.subdir,
66
  },