Datasets:
oliverkinch
commited on
Update README.md
Browse files
README.md
CHANGED
@@ -70,7 +70,13 @@ configs:
|
|
70 |
Each judgment/case consists of tabular data and a case-descriptive PDF. This dataset collects all these cases, with each sample describing a specific judgment/case.
|
71 |
|
72 |
The PDFs are anonymized to protect sensitive information. Therefore, each sample includes two text versions:
|
73 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
74 |
|
75 |
`text_anon` will be empty if no anonymization is detected in the PDF.
|
76 |
|
@@ -97,6 +103,7 @@ An example from the dataset looks as follows.
|
|
97 |
}
|
98 |
```
|
99 |
|
|
|
100 |
### Data Fields
|
101 |
|
102 |
- `case_id`: a `string` feature.
|
@@ -104,17 +111,23 @@ An example from the dataset looks as follows.
|
|
104 |
- `text`: a `string` feature.
|
105 |
- `text_anon`: a `string` feature.
|
106 |
- `text_len`: an `int` feature.
|
107 |
-
- `text_anon_len`: an `int` feature.
|
108 |
|
109 |
|
110 |
### Dataset Statistics
|
111 |
|
|
|
|
|
|
|
|
|
112 |
#### Number of samples
|
113 |
|
114 |
- 3919
|
115 |
|
116 |
#### PDF Text Length Distribution
|
117 |
|
|
|
|
|
118 |
- Minimum length: 192
|
119 |
- Maximum length: 2101736
|
120 |
|
@@ -122,7 +135,7 @@ An example from the dataset looks as follows.
|
|
122 |
|
123 |
## Potential Dataset Issues
|
124 |
|
125 |
-
See [open issues](https://github.com/oliverkinch/doms_databasen/issues)
|
126 |
|
127 |
|
128 |
## Dataset Creation
|
@@ -133,7 +146,7 @@ There are not many large-scale law datasets in Danish.
|
|
133 |
|
134 |
### Source Data
|
135 |
|
136 |
-
The dataset has been scraped from [Domsdatabasen](https://domsdatabasen.dk/)
|
137 |
|
138 |
## Additional Information
|
139 |
|
|
|
70 |
Each judgment/case consists of tabular data and a case-descriptive PDF. This dataset collects all these cases, with each sample describing a specific judgment/case.
|
71 |
|
72 |
The PDFs are anonymized to protect sensitive information. Therefore, each sample includes two text versions:
|
73 |
+
- `text_anon` (with anonymization tags: \<anonym\>"Some sensitive text"\</anonym\>).
|
74 |
+
- `text` (without anonymization tags).
|
75 |
+
|
76 |
+
`text_anon` is read with [Easyocr](https://github.com/JaidedAI/EasyOCR).
|
77 |
+
|
78 |
+
`text` is read with [Easyocr](https://github.com/JaidedAI/EasyOCR) or [Tika-python](https://github.com/chrismattmann/tika-python)
|
79 |
+
depending on the PDF and the anonymization method used.
|
80 |
|
81 |
`text_anon` will be empty if no anonymization is detected in the PDF.
|
82 |
|
|
|
103 |
}
|
104 |
```
|
105 |
|
106 |
+
|
107 |
### Data Fields
|
108 |
|
109 |
- `case_id`: a `string` feature.
|
|
|
111 |
- `text`: a `string` feature.
|
112 |
- `text_anon`: a `string` feature.
|
113 |
- `text_len`: an `int` feature.
|
114 |
+
- `text_anon_len`: an `int` feature.
|
115 |
|
116 |
|
117 |
### Dataset Statistics
|
118 |
|
119 |
+
#### Size of dataset
|
120 |
+
|
121 |
+
With the PDF texts being provided in two versions, `text` and `text_anon`, the total size of all PDF texts is approximately ~199//2 MB.
|
122 |
+
|
123 |
#### Number of samples
|
124 |
|
125 |
- 3919
|
126 |
|
127 |
#### PDF Text Length Distribution
|
128 |
|
129 |
+
Statistics based on `text`.
|
130 |
+
|
131 |
- Minimum length: 192
|
132 |
- Maximum length: 2101736
|
133 |
|
|
|
135 |
|
136 |
## Potential Dataset Issues
|
137 |
|
138 |
+
See [open issues](https://github.com/oliverkinch/doms_databasen/issues).
|
139 |
|
140 |
|
141 |
## Dataset Creation
|
|
|
146 |
|
147 |
### Source Data
|
148 |
|
149 |
+
The dataset has been scraped from [Domsdatabasen](https://domsdatabasen.dk/).
|
150 |
|
151 |
## Additional Information
|
152 |
|