Spaces:
Running
Running
Update app.py
Browse files
app.py
CHANGED
|
@@ -1,13 +1,14 @@
|
|
| 1 |
-
from typing import
|
| 2 |
import spacy
|
| 3 |
from spacy import displacy
|
| 4 |
from spacy.language import Language
|
| 5 |
import streamlit as st
|
| 6 |
from spacy_streamlit import visualize_parser
|
|
|
|
| 7 |
import base64
|
| 8 |
from PIL import Image
|
| 9 |
import deplacy
|
| 10 |
-
import graphviz
|
| 11 |
|
| 12 |
|
| 13 |
|
|
@@ -15,16 +16,16 @@ import graphviz
|
|
| 15 |
|
| 16 |
st.set_page_config(layout="wide")
|
| 17 |
|
| 18 |
-
st.
|
| 19 |
|
| 20 |
-
st.
|
| 21 |
|
| 22 |
-
st.
|
| 23 |
|
| 24 |
-
st.
|
| 25 |
-
spacy_model = st.
|
| 26 |
|
| 27 |
-
st.header("
|
| 28 |
text = st.text_area("", "Πλάτων ὁ Περικτιόνης τὸ γένος ἀνέφερεν εἰς Σόλωνα.")
|
| 29 |
|
| 30 |
|
|
@@ -47,7 +48,7 @@ def get_svg(svg: str, style: str = "", wrap: bool = True):
|
|
| 47 |
def visualize_parser(
|
| 48 |
doc: spacy.tokens.Doc,
|
| 49 |
*,
|
| 50 |
-
title: Optional[str] = "Dependency parse & part of speech",
|
| 51 |
key: Optional[str] = None,
|
| 52 |
) -> None:
|
| 53 |
"""Visualizer for dependency parses."""
|
|
@@ -61,7 +62,7 @@ def visualize_parser(
|
|
| 61 |
"collapse_punct": cols[1].checkbox(
|
| 62 |
"Collapse punct", value=True, key=f"{key}_parser_collapse_punct"
|
| 63 |
),
|
| 64 |
-
"compact": cols[3].checkbox("Compact mode", value=
|
| 65 |
}
|
| 66 |
docs = [span.as_doc() for span in doc.sents] if split_sents else [doc]
|
| 67 |
for sent in docs:
|
|
@@ -75,32 +76,11 @@ def visualize_parser(
|
|
| 75 |
|
| 76 |
visualize_parser(doc)
|
| 77 |
|
| 78 |
-
#graph_r = deplacy.render(doc)
|
| 79 |
-
|
| 80 |
-
#st.graphviz_chart(graph_r)
|
| 81 |
|
|
|
|
| 82 |
|
| 83 |
graph_dot = deplacy.dot(doc)
|
| 84 |
|
| 85 |
-
#graphviz.Source(deplacy.dot(doc))
|
| 86 |
-
|
| 87 |
st.graphviz_chart(graph_dot)
|
| 88 |
|
| 89 |
-
|
| 90 |
-
|
| 91 |
-
|
| 92 |
-
|
| 93 |
-
#st.sidebar.title("Model 2")
|
| 94 |
-
#spacy_model2 = st.sidebar.selectbox("Model 2", ["grc_ud_perseus_lg", "grc_ud_proiel_lg"])
|
| 95 |
-
|
| 96 |
-
#st.header("Text to analyze:")
|
| 97 |
-
#text = st.text_area("", "Πλάτων ὁ Περικτιόνης τὸ γένος ἀνέφερεν εἰς Σόλωνα.")
|
| 98 |
-
|
| 99 |
-
|
| 100 |
-
#nlp = spacy.load(spacy_model2)
|
| 101 |
-
#doc2 = nlp(text)
|
| 102 |
-
|
| 103 |
-
#visualize_parser(doc2)
|
| 104 |
-
|
| 105 |
-
#visualizers = ["pos", "dep"]
|
| 106 |
-
#spacy_streamlit.visualize(models, default_text,visualizers)
|
|
|
|
| 1 |
+
from typing import Optional
|
| 2 |
import spacy
|
| 3 |
from spacy import displacy
|
| 4 |
from spacy.language import Language
|
| 5 |
import streamlit as st
|
| 6 |
from spacy_streamlit import visualize_parser
|
| 7 |
+
from spacy_streamlit import visualize_tokens
|
| 8 |
import base64
|
| 9 |
from PIL import Image
|
| 10 |
import deplacy
|
| 11 |
+
#import graphviz
|
| 12 |
|
| 13 |
|
| 14 |
|
|
|
|
| 16 |
|
| 17 |
st.set_page_config(layout="wide")
|
| 18 |
|
| 19 |
+
st.image("logo.png", use_column_width=False, width=150)
|
| 20 |
|
| 21 |
+
st.title("Diogenet's Ancient Greek Syntax Analyzer")
|
| 22 |
|
| 23 |
+
st.markdown("Here you'll find four spaCy models for processing ancient Greek. They have been trained with the Universal Dependencies datasets *Perseus* and *Proiel*. We provide two types of models for each dataset. The '_lg' models were built with tok2vec pretrained embeddings and fasttext vectors, while the '_tfr' models have a transfomers layer. You can choose among models to compare their performance. More information about the models can be found in the [Huggingface Models Hub] (https://huggingface.co/Jacobo).")
|
| 24 |
|
| 25 |
+
st.header("Select a model:")
|
| 26 |
+
spacy_model = st.selectbox("", ["grc_ud_perseus_lg", "grc_ud_proiel_lg","grc_ud_perseus_trf"])
|
| 27 |
|
| 28 |
+
st.header("Enter text:")
|
| 29 |
text = st.text_area("", "Πλάτων ὁ Περικτιόνης τὸ γένος ἀνέφερεν εἰς Σόλωνα.")
|
| 30 |
|
| 31 |
|
|
|
|
| 48 |
def visualize_parser(
|
| 49 |
doc: spacy.tokens.Doc,
|
| 50 |
*,
|
| 51 |
+
title: Optional[str] = "Dependency parse & part of speech:",
|
| 52 |
key: Optional[str] = None,
|
| 53 |
) -> None:
|
| 54 |
"""Visualizer for dependency parses."""
|
|
|
|
| 62 |
"collapse_punct": cols[1].checkbox(
|
| 63 |
"Collapse punct", value=True, key=f"{key}_parser_collapse_punct"
|
| 64 |
),
|
| 65 |
+
"compact": cols[3].checkbox("Compact mode", value=False, key=f"{key}_parser_compact"),
|
| 66 |
}
|
| 67 |
docs = [span.as_doc() for span in doc.sents] if split_sents else [doc]
|
| 68 |
for sent in docs:
|
|
|
|
| 76 |
|
| 77 |
visualize_parser(doc)
|
| 78 |
|
|
|
|
|
|
|
|
|
|
| 79 |
|
| 80 |
+
st.header("Tree View:")
|
| 81 |
|
| 82 |
graph_dot = deplacy.dot(doc)
|
| 83 |
|
|
|
|
|
|
|
| 84 |
st.graphviz_chart(graph_dot)
|
| 85 |
|
| 86 |
+
visualize_tokens(doc, attrs=["text", "lemma_", "pos_", "dep_"], title="Table View:")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|