File size: 11,599 Bytes
8a5a6d8
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
import streamlit as st

from core.state import Metadata
from events.metadata import find_license_index
from events.metadata import LICENSES
from events.metadata import LICENSES_URL
from events.rai import handle_rai_change
from events.rai import RaiEvent

_INFO_TEXT = """This tab is the Responsible AI extension of Croissant. **Filling this tab is optional.**
        
More information on how to fill this part at: http://mlcommons.org/croissant/RAI/
"""


def render_rai_metadata():
    """Renders the `Metadata` view."""
    metadata: Metadata = st.session_state[Metadata]
    st.info(_INFO_TEXT, icon="💡")
    col1, col2 = st.columns([1, 1])
    with col1.expander("**Provenance**", expanded=True):
        with st.expander("**Data Collection**", expanded=False):
            key = "metadata-data-collection"
            st.text_area(
                label=("Explanation"),
                placeholder="Explain the key stages of the data collection process to improves understanding of potential users",
                key=key,
                value=metadata.data_collection,
                on_change=handle_rai_change,
                args=(RaiEvent.RAI_DATA_COLLECTION, metadata, key),
            )
            key = "metadata-data-collection-type"
            st.text_area(
                label=(
                    "Define the data collection type. Recommended values Recommended values: Surveys, Secondary Data analysis, Physical data collection, Direct measurement, Document analysis, Manual Human Curator, Software Collection, Experiments, Web Scraping, Web API, Focus groups, Self-reporting, Customer feedback data, User-generated content data, Passive Data Collection, Others"
                ),
                key=key,
                value=metadata.data_collection_type,
                on_change=handle_rai_change,
                args=(RaiEvent.RAI_DATA_COLLECTION_TYPE, metadata, key),
            )
            key = "metadata-data-collection-missing"
            st.text_area(
                label=("**Missing Data**."),
                key=key,
                placeholder="Description of missing data in structured/unstructured form",
                value=metadata.data_collection_missing_data,
                on_change=handle_rai_change,
                args=(RaiEvent.RAI_DATA_COLLECTION_MISSING_DATA, metadata, key),
            )
            key = "metadata-data-collection-raw"
            st.text_area(
                label=("**Raw Data**."),
                key=key,
                placeholder="Description of the raw data i.e. source of the data ",
                value=metadata.data_collection_raw_data,
                on_change=handle_rai_change,
                args=(RaiEvent.RAI_DATA_COLLECTION_RAW, metadata, key),
            )
        with st.expander("**Data Annotation**", expanded=False):
            key = "metadata-data-annotation-protocol"
            st.text_area(
                label=(
                    "**Protocol**. Description of annotations (labels, ratings) produced, including how these were created or authored -  Annotation Workforce Type, Annotation Characteristic(s), Annotation Description(s), Annotation Task(s), Annotation Distribution(s)"
                ),
                key=key,
                value=metadata.data_annotation_protocol,
                on_change=handle_rai_change,
                args=(RaiEvent.RAI_DATA_ANNOTATION_PROTOCOL, metadata, key),
            )
            key = "metadata-data-annotation-platform"
            st.text_area(
                label=(
                    "**Platform**. Platform, tool, or library used to collect annotations by human annotators"
                ),
                key=key,
                value=metadata.data_annotation_platform,
                on_change=handle_rai_change,
                args=(RaiEvent.RAI_DATA_ANNOTATION_PLATFORM, metadata, key),
            )
            key = "metadata-data-annotation-analysis"
            st.text_area(
                label=(
                    "**Analysis**. Considerations related to the process of converting the “raw” annotations into the labels that are ultimately packaged in a dataset - Uncertainty or disagreement between annotations on each instance as a signal in the dataset, analysis of systematic disagreements between annotators of different socio demographic group,  how the final dataset annotations will relate to individual annotator responses"
                ),
                key=key,
                value=metadata.data_annotation_analysis,
                on_change=handle_rai_change,
                args=(RaiEvent.RAI_DATA_ANNOTATION_ANALYSIS, metadata, key),
            )
            key = "metadata-data-annotation-demographics"
            st.text_area(
                label=(
                    "**Demographics**. List of demographics specifications about the annotators"
                ),
                key=key,
                value=metadata.annotator_demographics,
                on_change=handle_rai_change,
                args=(RaiEvent.RAI_DATA_ANNOTATION_DEMOGRAPHICS, metadata, key),
            )
            key = "metadata-data-annotation-tools"
            st.text_area(
                label=(
                    "**Tools**. List of software used for data annotation ( e.g. concept extraction, NER, and additional characteristics of the tools used for annotation to allow for replication or extension)  "
                ),
                key=key,
                value=metadata.machine_annotation_tools,
                on_change=handle_rai_change,
                args=(RaiEvent.RAI_DATA_ANNOTATION_TOOLS, metadata, key),
            )
            key = "metadata-data-annotation-per-item"
            st.text_area(
                label=(
                    "**Annotation per item**. Number of human labels per dataset item  "
                ),
                key=key,
                value=metadata.annotation_per_item,
                on_change=handle_rai_change,
                args=(RaiEvent.RAI_DATA_ANNOTATION_PER_ITEM, metadata, key),
            )
        with st.expander("**Data Preprocessing**", expanded=False):
            one_to_many_property(
                title="**Protocols**",
                metadata=metadata,
                attributes=metadata.data_preprocessing_protocol,
                key="metadata-data-preprocessing-protocol_",
                label="Description of data manipulation process if applicable ",
                event=RaiEvent.RAI_DATA_PREPROCESSING_PROTOCOL,
            )

            key = "metadata-data-manipulation-protocol"
            st.text_area(
                label=(
                    "**Manipulation**. Description of data manipulation process if applicable    "
                ),
                key=key,
                value=metadata.data_manipulation_protocol,
                on_change=handle_rai_change,
                args=(RaiEvent.RAI_DATA_MANIPULATION_PROTOCOL, metadata, key),
            )
            key = "metadata-data-imputation-protocol"
            st.text_area(
                label=(
                    "**Imputation**. Description of data imputation process if applicable  "
                ),
                key=key,
                value=metadata.data_imputation_protocol,
                on_change=handle_rai_change,
                args=(RaiEvent.RAI_DATA_IMPUTATION_PROTOCOL, metadata, key),
            )

    with col2.expander("**Data uses and social impact**", expanded=True):
        one_to_many_property(
            title="**Use cases**",
            metadata=metadata,
            attributes=metadata.data_use_cases,
            key="metadata-data-use-cases_",
            label="Dataset use case - training, testing, validation, development or production use, fine tuning, others (please specify), usage guidelines, recommended uses, etc.",
            event=RaiEvent.RAI_DATA_USE_CASES,
        )

        one_to_many_property(
            title="**Data biases**",
            metadata=metadata,
            attributes=metadata.data_biases,
            key="metadata-data-biases_",
            label="**Data biases**. Involves understanding the potential risks associated  with data usage and to prevent unintended and potentially harmful consequences that may arise from using models trained on or evaluated with the respective data",
            event=RaiEvent.RAI_DATA_BIAS,
        )

        one_to_many_property(
            title="**Personal and sensitive information**",
            metadata=metadata,
            attributes=metadata.personal_sensitive_information,
            key="metadata-personal-sensitive-information_",
            label="Personal and sensitive information, if contained within the dataset, can play an important role in the mitigation of any risks and the responsible use of the datasets",
            event=RaiEvent.RAI_SENSITIVE,
        )

        key = "metadata-social-impact"
        st.text_area(
            label=("**Social impact**. Discussion of social impact, if applicable"),
            key=key,
            value=metadata.data_social_impact,
            on_change=handle_rai_change,
            args=(RaiEvent.RAI_DATA_SOCIAL_IMPACT, metadata, key),
        )

        one_to_many_property(
            "**Data limitations**",
            metadata,
            metadata.data_limitations,
            "metadata-data-limitations_",
            "Known limitations - Data generalization limits (e.g related to data distribution, data quality issues, or data sources) and on-recommended uses.",
            RaiEvent.RAI_DATA_LIMITATION,
        )

        key = "metadata-data-maintenance"
        st.text_area(
            label=(
                "**Data release maintenance**. Versioning information in terms of the updating timeframe, the maintainers, and the deprecation policies. "
            ),
            key=key,
            value=metadata.data_release_maintenance_plan,
            on_change=handle_rai_change,
            args=(RaiEvent.RAI_MAINTENANCE, metadata, key),
        )


def one_to_many_property(
    title: str, metadata: Metadata, attributes, key: str, label: str, event: str
):
    """Generates a one to many cardinality property. Attributes should be empty, have one element or being a list of elements"""
    with st.expander(title, expanded=True):
        if attributes:
            if not isinstance(attributes, list):
                attributes = [attributes]
            for index, single_attribute in enumerate(attributes):
                key = key + str(index)
                st.text_area(
                    label=(label),
                    key=key,
                    value=single_attribute,
                    on_change=handle_rai_change,
                    args=(event, metadata, key, index),
                )
        else:
            key = key + "0"
            st.text_area(
                label=(label),
                key=key,
                on_change=handle_rai_change,
                args=(event, metadata, key),
            )
        add, remove = st.columns(2)
        with add:
            if st.button("+ add", key=key + "add"):
                if attributes:
                    attributes.append("")
                    st.rerun()
                else:
                    attributes = []
                    attributes.append("")
                    st.rerun()
        with remove:
            if st.button("- remove", key=key + "remove"):
                if attributes:
                    attributes.pop()
                    st.rerun()