Datasets:

wangdx25 commited on
Commit
52b6f3d
·
verified ·
1 Parent(s): 11cec7b

Add TsFile (converted from williamgilpin/dysts)

Browse files
.gitattributes CHANGED
@@ -58,3 +58,9 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
58
  # Video files - compressed
59
  *.mp4 filter=lfs diff=lfs merge=lfs -text
60
  *.webm filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
 
 
58
  # Video files - compressed
59
  *.mp4 filter=lfs diff=lfs merge=lfs -text
60
  *.webm filter=lfs diff=lfs merge=lfs -text
61
+ dysts_coarse_1.tsfile filter=lfs diff=lfs merge=lfs -text
62
+ dysts_coarse_2.tsfile filter=lfs diff=lfs merge=lfs -text
63
+ dysts_fine_1.tsfile filter=lfs diff=lfs merge=lfs -text
64
+ dysts_fine_2.tsfile filter=lfs diff=lfs merge=lfs -text
65
+ dysts_medium_1.tsfile filter=lfs diff=lfs merge=lfs -text
66
+ dysts_medium_2.tsfile filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,148 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ pretty_name: "Dysts Chaotic Dynamical-System Trajectories (TsFile)"
3
+ modality: timeseries
4
+ license: cc-by-4.0
5
+ authors: "William Gilpin"
6
+ task_categories:
7
+ - time-series-forecasting
8
+ size_categories:
9
+ - 1M<n<10M
10
+ tags:
11
+ - tsfile
12
+ - timeseries
13
+ - modality:timeseries
14
+ - format:tsfile
15
+ - dynamical-systems
16
+ - chaotic-systems
17
+ - simulation
18
+ configs:
19
+ - config_name: coarse
20
+ data_files:
21
+ - split: coarse
22
+ path: "dysts_coarse_*.tsfile"
23
+ - config_name: medium
24
+ data_files:
25
+ - split: medium
26
+ path: "dysts_medium_*.tsfile"
27
+ - config_name: fine
28
+ data_files:
29
+ - split: fine
30
+ path: "dysts_fine_*.tsfile"
31
+ ---
32
+
33
+ # Dysts Chaotic Dynamical-System Trajectories (TsFile)
34
+
35
+ This dataset is the Apache TsFile conversion of
36
+ [`williamgilpin/dysts`](https://huggingface.co/datasets/williamgilpin/dysts),
37
+ a collection of simulated trajectories from named chaotic and nonlinear
38
+ dynamical systems. See the source papers [arXiv:2110.05266](https://arxiv.org/abs/2110.05266)
39
+ and [arXiv:2303.08011](https://arxiv.org/abs/2303.08011).
40
+
41
+ Modalities: Time-series.
42
+
43
+ ## Overview
44
+
45
+ - Source revision: `5d305abb3c9010e36ca6a3e6c967b72932f6cca9`
46
+ - Source files: `coarse/*.csv`, `medium/*.csv`, and `fine/*.csv`
47
+ - Trajectories/files: 352 total (117 coarse, 118 medium, 117 fine)
48
+ - Points per trajectory: 10,000
49
+ - Converted rows: 3,520,000 total (1,170,000 coarse, 1,180,000 medium,
50
+ 1,170,000 fine)
51
+ - License: `cc-by-4.0`
52
+
53
+ Each CSV contains a `time` column and a variable number of numeric channels.
54
+ The source channel names are commonly `x0`, ` x1`, ...; the leading whitespace
55
+ is normalized during conversion. Different systems may have different channel
56
+ counts, so each output table uses a union schema.
57
+
58
+ ## TsFile files and schema
59
+
60
+ The converter emits one logical table per source resolution. Each logical
61
+ table is split into two TsFile shards by the import tool:
62
+
63
+ | File | Table | Rows | Size (bytes) |
64
+ |---|---|---:|---:|
65
+ | `dysts_coarse_1.tsfile` | `dysts_coarse` | 1,048,576 | 29,477,332 |
66
+ | `dysts_coarse_2.tsfile` | `dysts_coarse` | 121,424 | 3,581,515 |
67
+ | `dysts_medium_1.tsfile` | `dysts_medium` | 1,048,576 | 30,225,699 |
68
+ | `dysts_medium_2.tsfile` | `dysts_medium` | 131,424 | 3,893,846 |
69
+ | `dysts_fine_1.tsfile` | `dysts_fine` | 1,048,576 | 30,972,948 |
70
+ | `dysts_fine_2.tsfile` | `dysts_fine` | 121,424 | 3,527,556 |
71
+
72
+ Every table has the following common columns:
73
+
74
+ | Column | Role | Type | Meaning |
75
+ |---|---|---|---|
76
+ | `Time` | TIME | INT64 (ms) | Rounded source `time * 1000`; source values run from 0 to 9,999 |
77
+ | `granularity` | TAG | STRING | `coarse`, `medium`, or `fine` |
78
+ | `system` | TAG | STRING | System name parsed from the source filename |
79
+ | `x0` through `x9` | FIELD | DOUBLE | Numeric trajectory channels; unavailable channels are null |
80
+
81
+ The actual number of populated channel fields depends on the system. The
82
+ union schema keeps the maximum observed channel names (`x0` through `x9`) in each
83
+ resolution table.
84
+
85
+ ## Conversion notes
86
+
87
+ - Source `time` is interpreted as seconds and converted to integer-millisecond
88
+ `Time` by rounding. It is not retained as a second field.
89
+ - The filename suffix (`_coarse`, `_medium`, `_fine`) supplies the
90
+ `granularity` TAG; the remaining filename stem supplies `system`.
91
+ - Channel names are trimmed and sanitized to TsFile-safe identifiers. Numeric
92
+ values are copied without scaling. Missing channels in a system are null;
93
+ no values are imputed and no trajectory rows are dropped.
94
+ - Files are processed deterministically and sorted by `granularity`, `system`,
95
+ and `Time`. Each source system has unique `(system, Time)` keys within its
96
+ resolution.
97
+
98
+ ## Read example
99
+
100
+ ```python
101
+ from tsfile import TsFileReader
102
+
103
+ path = "dysts_coarse_1.tsfile"
104
+ with TsFileReader(path) as reader:
105
+ with reader.query_table(
106
+ "dysts_coarse",
107
+ ["x0", "x1", "x2"],
108
+ batch_size=4096,
109
+ ) as result:
110
+ batch = result.read_arrow_batch()
111
+ if batch is not None:
112
+ print(batch.to_pandas().head())
113
+ ```
114
+
115
+ ## Source & license
116
+
117
+ - Original dataset: https://huggingface.co/datasets/williamgilpin/dysts
118
+ - Author / publisher: William Gilpin
119
+ - Papers: https://arxiv.org/abs/2110.05266 and https://arxiv.org/abs/2303.08011
120
+ - License: CC-BY-4.0
121
+
122
+ ## Usage
123
+
124
+ Install the Apache TsFile Python SDK (`pip install tsfile`) and read a converted file:
125
+
126
+ ```python
127
+ from pathlib import Path
128
+ from tsfile import TsFileReader
129
+
130
+ path = Path("dysts_coarse_1.tsfile")
131
+ with TsFileReader(str(path)) as reader:
132
+ schemas = reader.get_all_table_schemas()
133
+ print("tables:", list(schemas))
134
+ table_name = next(iter(schemas))
135
+ table = schemas[table_name]
136
+ columns = [column.get_column_name() for column in table.get_columns()]
137
+ print("columns:", columns)
138
+ field_names = [
139
+ column.get_column_name()
140
+ for column in table.get_columns()
141
+ if column.get_column_name() not in {"Time", "time"}
142
+ ]
143
+ if field_names:
144
+ with reader.query_table(table_name, field_names[:3], batch_size=1024) as result:
145
+ batch = result.read_arrow_batch()
146
+ if batch is not None:
147
+ print(batch.to_pandas().head())
148
+ ```
dysts_coarse_1.tsfile ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c03de3da0fc04a2874378690492bd3d35adbf67f79fc811b627a6b8d426b9f4f
3
+ size 29477332
dysts_coarse_2.tsfile ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a14ac2904d68815e2b1bf3b2c382878996adee377c7ba8f64f842875bda777e8
3
+ size 3581515
dysts_fine_1.tsfile ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4c7e69fb7cc1513e0eca7c832ee4f0ffa6f031c47d06cabb89a4d263ea65472b
3
+ size 30972948
dysts_fine_2.tsfile ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7a069aba5e608bac277d6d358e294400613f7f0a780f1bfee2eaf3d712317278
3
+ size 3527556
dysts_medium_1.tsfile ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d6476e5b9b1d7d65d1fa7b8578a3c0b55fd8f59258e9e009e27fdafa1abbee68
3
+ size 30225699
dysts_medium_2.tsfile ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d273f637e53ff68374286f017b9b3617d7e8c250ee537e8d0fc2765f43439c41
3
+ size 3893846