Skip to content

edsnlp.connectors.omop

OmopConnector

Bases: object

[summary]

PARAMETER DESCRIPTION
nlp

spaCy language object.

TYPE: Language

start_char

Name of the column containing the start character index of the entity, by default "start_char"

TYPE: str, optional

end_char

Name of the column containing the end character index of the entity, by default "end_char"

TYPE: str, optional

Source code in edsnlp/connectors/omop.py 
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
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
class OmopConnector(object):
    """
    [summary]

    Parameters
    ----------
    nlp : Language
        spaCy language object.
    start_char : str, optional
        Name of the column containing the start character index of the entity,
        by default "start_char"
    end_char : str, optional
        Name of the column containing the end character index of the entity,
        by default "end_char"
    """

    def __init__(
        self,
        nlp: Language,
        start_char: str = "start_char",
        end_char: str = "end_char",
    ):

        self.start_char = start_char
        self.end_char = end_char

        self.nlp = nlp

    def preprocess(
        self, note: pd.DataFrame, note_nlp: pd.DataFrame
    ) -> Tuple[pd.DataFrame, pd.DataFrame]:
        """
        Preprocess the input OMOP tables: modification of the column names.

        Parameters
        ----------
        note : pd.DataFrame
            OMOP `note` table.
        note_nlp : pd.DataFrame
            OMOP `note_nlp` table.

        Returns
        -------
        note : pd.DataFrame
            OMOP `note` table.
        note_nlp : pd.DataFrame
            OMOP `note_nlp` table.
        """

        note_nlp = note_nlp.rename(
            columns={
                self.start_char: "start_char",
                self.end_char: "end_char",
            }
        )

        return note, note_nlp

    def postprocess(
        self, note: pd.DataFrame, note_nlp: pd.DataFrame
    ) -> Tuple[pd.DataFrame, pd.DataFrame]:
        """
        Postprocess the input OMOP tables: modification of the column names.

        Parameters
        ----------
        note : pd.DataFrame
            OMOP `note` table.
        note_nlp : pd.DataFrame
            OMOP `note_nlp` table.

        Returns
        -------
        note : pd.DataFrame
            OMOP `note` table.
        note_nlp : pd.DataFrame
            OMOP `note_nlp` table.
        """

        note_nlp = note_nlp.rename(
            columns={
                "start_char": self.start_char,
                "end_char": self.end_char,
            }
        )

        return note, note_nlp

    def omop2docs(
        self,
        note: pd.DataFrame,
        note_nlp: pd.DataFrame,
        extensions: Optional[List[str]] = None,
    ) -> List[Doc]:
        """
        Transforms OMOP tables to a list of spaCy documents.

        Parameters
        ----------
        note : pd.DataFrame
            OMOP `note` table.
        note_nlp : pd.DataFrame
            OMOP `note_nlp` table.
        extensions : Optional[List[str]], optional
            Extensions to keep, by default None

        Returns
        -------
        List[Doc]
            List of spaCy documents.
        """
        note, note_nlp = self.preprocess(note, note_nlp)
        return omop2docs(note, note_nlp, self.nlp, extensions)

    def docs2omop(
        self,
        docs: List[Doc],
        extensions: Optional[List[str]] = None,
    ) -> Tuple[pd.DataFrame, pd.DataFrame]:
        """
        Transforms a list of spaCy documents to a pair of OMOP tables.

        Parameters
        ----------
        docs : List[Doc]
            List of spaCy documents.
        extensions : Optional[List[str]], optional
            Extensions to keep, by default None

        Returns
        -------
        note : pd.DataFrame
            OMOP `note` table.
        note_nlp : pd.DataFrame
            OMOP `note_nlp` table.
        """
        note, note_nlp = docs2omop(docs, extensions=extensions)
        note, note_nlp = self.postprocess(note, note_nlp)
        return note, note_nlp

start_char = start_char instance-attribute

end_char = end_char instance-attribute

nlp = nlp instance-attribute

__init__(nlp, start_char='start_char', end_char='end_char')

Source code in edsnlp/connectors/omop.py 
201
202
203
204
205
206
207
208
209
210
211
def __init__(
    self,
    nlp: Language,
    start_char: str = "start_char",
    end_char: str = "end_char",
):

    self.start_char = start_char
    self.end_char = end_char

    self.nlp = nlp

preprocess(note, note_nlp)

Preprocess the input OMOP tables: modification of the column names.

PARAMETER DESCRIPTION
note

OMOP note table.

TYPE: pd.DataFrame

note_nlp

OMOP note_nlp table.

TYPE: pd.DataFrame

RETURNS DESCRIPTION
note

OMOP note table.

TYPE: pd.DataFrame

note_nlp

OMOP note_nlp table.

TYPE: pd.DataFrame

Source code in edsnlp/connectors/omop.py 
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
def preprocess(
    self, note: pd.DataFrame, note_nlp: pd.DataFrame
) -> Tuple[pd.DataFrame, pd.DataFrame]:
    """
    Preprocess the input OMOP tables: modification of the column names.

    Parameters
    ----------
    note : pd.DataFrame
        OMOP `note` table.
    note_nlp : pd.DataFrame
        OMOP `note_nlp` table.

    Returns
    -------
    note : pd.DataFrame
        OMOP `note` table.
    note_nlp : pd.DataFrame
        OMOP `note_nlp` table.
    """

    note_nlp = note_nlp.rename(
        columns={
            self.start_char: "start_char",
            self.end_char: "end_char",
        }
    )

    return note, note_nlp

postprocess(note, note_nlp)

Postprocess the input OMOP tables: modification of the column names.

PARAMETER DESCRIPTION
note

OMOP note table.

TYPE: pd.DataFrame

note_nlp

OMOP note_nlp table.

TYPE: pd.DataFrame

RETURNS DESCRIPTION
note

OMOP note table.

TYPE: pd.DataFrame

note_nlp

OMOP note_nlp table.

TYPE: pd.DataFrame

Source code in edsnlp/connectors/omop.py 
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
def postprocess(
    self, note: pd.DataFrame, note_nlp: pd.DataFrame
) -> Tuple[pd.DataFrame, pd.DataFrame]:
    """
    Postprocess the input OMOP tables: modification of the column names.

    Parameters
    ----------
    note : pd.DataFrame
        OMOP `note` table.
    note_nlp : pd.DataFrame
        OMOP `note_nlp` table.

    Returns
    -------
    note : pd.DataFrame
        OMOP `note` table.
    note_nlp : pd.DataFrame
        OMOP `note_nlp` table.
    """

    note_nlp = note_nlp.rename(
        columns={
            "start_char": self.start_char,
            "end_char": self.end_char,
        }
    )

    return note, note_nlp

omop2docs(note, note_nlp, extensions=None)

Transforms OMOP tables to a list of spaCy documents.

PARAMETER DESCRIPTION
note

OMOP note table.

TYPE: pd.DataFrame

note_nlp

OMOP note_nlp table.

TYPE: pd.DataFrame

extensions

Extensions to keep, by default None

TYPE: Optional[List[str]], optional DEFAULT: None

RETURNS DESCRIPTION
List[Doc]

List of spaCy documents.
Source code in edsnlp/connectors/omop.py 
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
def omop2docs(
    self,
    note: pd.DataFrame,
    note_nlp: pd.DataFrame,
    extensions: Optional[List[str]] = None,
) -> List[Doc]:
    """
    Transforms OMOP tables to a list of spaCy documents.

    Parameters
    ----------
    note : pd.DataFrame
        OMOP `note` table.
    note_nlp : pd.DataFrame
        OMOP `note_nlp` table.
    extensions : Optional[List[str]], optional
        Extensions to keep, by default None

    Returns
    -------
    List[Doc]
        List of spaCy documents.
    """
    note, note_nlp = self.preprocess(note, note_nlp)
    return omop2docs(note, note_nlp, self.nlp, extensions)

docs2omop(docs, extensions=None)

Transforms a list of spaCy documents to a pair of OMOP tables.

PARAMETER DESCRIPTION
docs

List of spaCy documents.

TYPE: List[Doc]

extensions

Extensions to keep, by default None

TYPE: Optional[List[str]], optional DEFAULT: None

RETURNS DESCRIPTION
note

OMOP note table.

TYPE: pd.DataFrame

note_nlp

OMOP note_nlp table.

TYPE: pd.DataFrame

Source code in edsnlp/connectors/omop.py 
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
def docs2omop(
    self,
    docs: List[Doc],
    extensions: Optional[List[str]] = None,
) -> Tuple[pd.DataFrame, pd.DataFrame]:
    """
    Transforms a list of spaCy documents to a pair of OMOP tables.

    Parameters
    ----------
    docs : List[Doc]
        List of spaCy documents.
    extensions : Optional[List[str]], optional
        Extensions to keep, by default None

    Returns
    -------
    note : pd.DataFrame
        OMOP `note` table.
    note_nlp : pd.DataFrame
        OMOP `note_nlp` table.
    """
    note, note_nlp = docs2omop(docs, extensions=extensions)
    note, note_nlp = self.postprocess(note, note_nlp)
    return note, note_nlp

omop2docs(note, note_nlp, nlp, extensions=None)

Transforms an OMOP-formatted pair of dataframes into a list of documents.

PARAMETER DESCRIPTION
note

The OMOP note table.

TYPE: pd.DataFrame

note_nlp

The OMOP note_nlp table

TYPE: pd.DataFrame

nlp

spaCy language object.

TYPE: Language

extensions

Extensions to keep, by default None

TYPE: Optional[List[str]], optional DEFAULT: None

RETURNS DESCRIPTION
List[Doc]

List of spaCy documents
Source code in edsnlp/connectors/omop.py 
 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
def omop2docs(
    note: pd.DataFrame,
    note_nlp: pd.DataFrame,
    nlp: Language,
    extensions: Optional[List[str]] = None,
) -> List[Doc]:
    """
    Transforms an OMOP-formatted pair of dataframes into a list of documents.

    Parameters
    ----------
    note : pd.DataFrame
        The OMOP `note` table.
    note_nlp : pd.DataFrame
        The OMOP `note_nlp` table
    nlp : Language
        spaCy language object.
    extensions : Optional[List[str]], optional
        Extensions to keep, by default None

    Returns
    -------
    List[Doc] :
        List of spaCy documents
    """

    note = note.copy()
    note_nlp = note_nlp.copy()

    extensions = extensions or []

    def row2ent(row):
        d = dict(
            start_char=row.start_char,
            end_char=row.end_char,
            label=row.get("note_nlp_source_value"),
            extensions={ext: row.get(ext) for ext in extensions},
        )

        return d

    # Create entities
    note_nlp["ents"] = note_nlp.apply(row2ent, axis=1)

    note_nlp = note_nlp.groupby("note_id", as_index=False)["ents"].agg(list)

    note = note.merge(note_nlp, on=["note_id"], how="left")

    # Generate documents
    note["doc"] = note.note_text.apply(nlp)

    # Process documents
    for _, row in note.iterrows():

        doc = row.doc
        doc._.note_id = row.note_id
        doc._.note_datetime = row.get("note_datetime")

        ents = []

        if not isinstance(row.ents, list):
            continue

        for ent in row.ents:

            span = doc.char_span(
                ent["start_char"],
                ent["end_char"],
                ent["label"],
                alignment_mode="expand",
            )

            for k, v in ent["extensions"].items():
                setattr(span._, k, v)

            ents.append(span)

            if span.label_ not in doc.spans:
                doc.spans[span.label_] = [span]
            else:
                doc.spans[span.label_].append(span)

        ents, discarded = filter_spans(ents, return_discarded=True)

        doc.ents = ents

        if "discarded" not in doc.spans:
            doc.spans["discarded"] = []
        doc.spans["discarded"].extend(discarded)

    return list(note.doc)

docs2omop(docs, extensions=None)

Transforms a list of spaCy docs to a pair of OMOP tables.

PARAMETER DESCRIPTION
docs

List of documents to transform.

TYPE: List[Doc]

extensions

Extensions to keep, by default None

TYPE: Optional[List[str]], optional DEFAULT: None

RETURNS DESCRIPTION
Tuple[pd.DataFrame, pd.DataFrame]

Pair of OMOP tables (note and note_nlp)
Source code in edsnlp/connectors/omop.py 
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
def docs2omop(
    docs: List[Doc],
    extensions: Optional[List[str]] = None,
) -> Tuple[pd.DataFrame, pd.DataFrame]:
    """
    Transforms a list of spaCy docs to a pair of OMOP tables.

    Parameters
    ----------
    docs : List[Doc]
        List of documents to transform.
    extensions : Optional[List[str]], optional
        Extensions to keep, by default None

    Returns
    -------
    Tuple[pd.DataFrame, pd.DataFrame]
        Pair of OMOP tables (`note` and `note_nlp`)
    """

    df = pd.DataFrame(dict(doc=docs))

    df["note_text"] = df.doc.apply(lambda doc: doc.text)
    df["note_id"] = df.doc.apply(lambda doc: doc._.note_id)
    df["note_datetime"] = df.doc.apply(lambda doc: doc._.note_datetime)

    if df.note_id.isna().any():
        df["note_id"] = range(len(df))

    df["ents"] = df.doc.apply(lambda doc: list(doc.ents))
    df["ents"] += df.doc.apply(lambda doc: list(doc.spans["discarded"]))

    note = df[["note_id", "note_text", "note_datetime"]]

    df = df[["note_id", "ents"]].explode("ents")

    extensions = extensions or []

    def ent2dict(
        ent: Span,
    ) -> Dict[str, Any]:

        d = dict(
            start_char=ent.start_char,
            end_char=ent.end_char,
            note_nlp_source_value=ent.label_,
            lexical_variant=ent.text,
            # normalized_variant=ent._.normalized.text,
        )

        for ext in extensions:
            d[ext] = getattr(ent._, ext)

        return d

    df["ents"] = df.ents.apply(ent2dict)

    columns = [
        "start_char",
        "end_char",
        "note_nlp_source_value",
        "lexical_variant",
        # "normalized_variant",
    ]
    columns += extensions

    df[columns] = df.ents.apply(pd.Series)

    df["term_modifiers"] = ""

    for i, ext in enumerate(extensions):
        if i > 0:
            df.term_modifiers += ";"
        df.term_modifiers += ext + "=" + df[ext].astype(str)

    df["note_nlp_id"] = range(len(df))

    note_nlp = df[["note_nlp_id", "note_id"] + columns]

    return note, note_nlp