Update README.md
Browse files
README.md
CHANGED
@@ -47,22 +47,6 @@ dataset_info:
|
|
47 |
- name: other
|
48 |
sequence: int64
|
49 |
configs:
|
50 |
-
- config_name: large_context
|
51 |
-
data_files:
|
52 |
-
- split: test
|
53 |
-
path: data/large_context/*
|
54 |
-
- config_name: medium_context
|
55 |
-
data_files:
|
56 |
-
- split: test
|
57 |
-
path: data/medium_context/*
|
58 |
-
- config_name: small_context
|
59 |
-
data_files:
|
60 |
-
- split: test
|
61 |
-
path: data/small_context/*
|
62 |
-
- config_name: huge_context
|
63 |
-
data_files:
|
64 |
-
- split: test
|
65 |
-
path: data/huge_context/*
|
66 |
- config_name: large_context
|
67 |
data_files:
|
68 |
- split: test
|
@@ -87,7 +71,14 @@ configs:
|
|
87 |
```
|
88 |
from datasets import load_dataset
|
89 |
|
90 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
91 |
```
|
92 |
## Data Point Structure
|
93 |
|
@@ -97,12 +88,12 @@ ds = load_dataset('JetBrains-Research/lca-codegen-huge', split='test')
|
|
97 |
* `filename` β filepath to the completion file
|
98 |
* `content` β content of the completion file
|
99 |
* `completion_lines` β dictionary where keys are classes of lines and values are a list of integers (numbers of lines to complete). The classes are:
|
100 |
-
* `committed` β line contains at least one function or class that was declared in the committed files from `commit_hash`
|
101 |
-
* `inproject` β line contains at least one function or class that was declared in the project (excluding previous)
|
102 |
-
* `infile` β line contains at least one function or class that was declared in the completion file (excluding previous)
|
103 |
-
* `common` β line contains at least one function or class that was classified to be common, e.g., `main`, `get`, etc (excluding previous)
|
104 |
-
* `non_informative` β line that was classified to be non-informative, e.g. too short, contains comments, etc
|
105 |
-
* `random` β randomly sampled from the rest of the lines
|
106 |
* `repo_snapshot` β dictionary with a snapshot of the repository before the commit. Has the same structure as `completion_file`, but filenames and contents are orginized as lists.
|
107 |
* `completion_lines_raw` β the same as `completion_lines`, but before sampling.
|
108 |
|
@@ -112,33 +103,58 @@ To collect the data, we cloned repositories from GitHub where the main language
|
|
112 |
The completion file for each data point is a `.py` file that was added to the repository in a commit.
|
113 |
The state of the repository before this commit is the repo snapshot.
|
114 |
|
115 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
116 |
|
117 |
-
|
|
|
|
|
|
|
|
|
|
|
118 |
|
119 |
-
* Number of datapoints: 296
|
120 |
-
* Number of repositories: 75
|
121 |
-
* Number of commits: 252
|
122 |
|
123 |
### Completion File
|
124 |
-
|
125 |
-
|
126 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
127 |
|
128 |
-
### Repository Snapshot
|
129 |
-
* `.py` files: <u>median 261</u>, from 47 to 5227
|
130 |
-
* non `.py` files: <u>median 262</u>, from 24 to 7687
|
131 |
-
* `.py` lines: <u>median 49811</u>
|
132 |
-
* non `.py` lines: <u>median 60163</u>
|
133 |
|
134 |
### Line Counts:
|
135 |
-
*
|
136 |
-
|
137 |
-
|
138 |
-
|
139 |
-
|
140 |
-
|
141 |
-
|
142 |
|
143 |
## Scores
|
144 |
[HF Space](https://huggingface.co/spaces/JetBrains-Research/long-code-arena)
|
|
|
47 |
- name: other
|
48 |
sequence: int64
|
49 |
configs:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
50 |
- config_name: large_context
|
51 |
data_files:
|
52 |
- split: test
|
|
|
71 |
```
|
72 |
from datasets import load_dataset
|
73 |
|
74 |
+
config_names = [
|
75 |
+
'small_context',
|
76 |
+
'medium_context',
|
77 |
+
'large_context',
|
78 |
+
'huge_context'
|
79 |
+
]
|
80 |
+
|
81 |
+
ds = load_dataset('JetBrains-Research/lca-code-completion', config_name, split='test')
|
82 |
```
|
83 |
## Data Point Structure
|
84 |
|
|
|
88 |
* `filename` β filepath to the completion file
|
89 |
* `content` β content of the completion file
|
90 |
* `completion_lines` β dictionary where keys are classes of lines and values are a list of integers (numbers of lines to complete). The classes are:
|
91 |
+
* `committed` β line contains at least one function or class that was declared in the committed files from `commit_hash`
|
92 |
+
* `inproject` β line contains at least one function or class that was declared in the project (excluding previous)
|
93 |
+
* `infile` β line contains at least one function or class that was declared in the completion file (excluding previous)
|
94 |
+
* `common` β line contains at least one function or class that was classified to be common, e.g., `main`, `get`, etc (excluding previous)
|
95 |
+
* `non_informative` β line that was classified to be non-informative, e.g. too short, contains comments, etc
|
96 |
+
* `random` β randomly sampled from the rest of the lines
|
97 |
* `repo_snapshot` β dictionary with a snapshot of the repository before the commit. Has the same structure as `completion_file`, but filenames and contents are orginized as lists.
|
98 |
* `completion_lines_raw` β the same as `completion_lines`, but before sampling.
|
99 |
|
|
|
103 |
The completion file for each data point is a `.py` file that was added to the repository in a commit.
|
104 |
The state of the repository before this commit is the repo snapshot.
|
105 |
|
106 |
+
The dataset configurations are based on the number of characters in `.py` files from the repository snapshot:
|
107 |
+
- `small_context` β less than 48K characters;
|
108 |
+
- `medium_context` β from 48K to 192K characters;
|
109 |
+
- `large_context` β from 192K to 768K characters;
|
110 |
+
- `huge_context` β more than 768K characters.
|
111 |
+
|
112 |
+
## Datasets Stats
|
113 |
|
114 |
+
| Dataset | Number of datapoints | Number of repositories | Number of commits |
|
115 |
+
|:-----------------:|:--------------------:|:----------------------:|:-----------------:|
|
116 |
+
| `small_context` | 144 | 46 | 63 |
|
117 |
+
| `medium_context` | 224 | 80 | 175 |
|
118 |
+
| `large_context` | 270 | 75 | 219 |
|
119 |
+
| `huge_context` | 296 | 75 | 252 |
|
120 |
|
|
|
|
|
|
|
121 |
|
122 |
### Completion File
|
123 |
+
| Dataset | Completion file lines, min | Completion file lines, max | Completion file lines, median |
|
124 |
+
|:------------------:|:--------------------------:|:--------------------------:|:------------------------------:|
|
125 |
+
| `small_context` | 201 | 1916 | 310.5 |
|
126 |
+
| `medium_context` | 200 | 1648 | 310.0 |
|
127 |
+
| `large_context` | 200 | 1694 | 278.0 |
|
128 |
+
| `huge_context` | 200 | 1877 | 313.5 |
|
129 |
+
|
130 |
+
|
131 |
+
|
132 |
+
### Repository Snapshot `.py` files
|
133 |
+
| Dataset | Context py files number, min | Context py files number, max | Context py files number, median | Context py lines, median |
|
134 |
+
|:------------------:|:----------------------------:|:----------------------------:|:--------------------------------:|:-------------------------:|
|
135 |
+
| `small_context` | 0 | 52 | 4.0 | 128.0 |
|
136 |
+
| `medium_context` | 3 | 117 | 34.0 | 3786.0 |
|
137 |
+
| `large_context` | 3 | 255 | 84.0 | 15466.5 |
|
138 |
+
| `huge_context` | 47 | 5227 | 261.0 | 49811.0 |
|
139 |
+
|
140 |
+
|
141 |
+
### Repository Snapshot non `.py` files
|
142 |
+
| Dataset | Context non-py files number, min | Context non-py files number, max | Context non-py files number, median | Context non-py lines, median |
|
143 |
+
|:------------------:|:---------------------------------:|:---------------------------------:|:-----------------------------------:|:-----------------------------:|
|
144 |
+
| `small_context` | 1 | 1044 | 19.5 | 1227.0 |
|
145 |
+
| `medium_context` | 3 | 3977 | 64.5 | 9735.0 |
|
146 |
+
| `large_context` | 8 | 2174 | 155.0 | 18759.0 |
|
147 |
+
| `huge_context` | 24 | 7687 | 262.0 | 60163.0 |
|
148 |
|
|
|
|
|
|
|
|
|
|
|
149 |
|
150 |
### Line Counts:
|
151 |
+
| Dataset | *infile* | *inproject* | *common* | *commited* | *non-informative* | *random* | **all** |
|
152 |
+
|:------------------:|:--------:|:-----------:|:--------:|:----------:|:-----------------:|:--------:|:-----:|
|
153 |
+
| `small_context` | 1430 | 95 | 500 | 1426 | 532 | 703 | **4686** |
|
154 |
+
| `medium_context` | 2224 | 2236 | 779 | 1495 | 858 | 1084 | **8676** |
|
155 |
+
| `large_context` | 2691 | 2595 | 693 | 1322 | 1019 | 1311 | **9631** |
|
156 |
+
| `huge_context` | 2608 | 2901 | 692 | 1019 | 1164 | 1426 | **9810** |
|
157 |
+
|
158 |
|
159 |
## Scores
|
160 |
[HF Space](https://huggingface.co/spaces/JetBrains-Research/long-code-arena)
|