Spaces:
Sleeping
Sleeping
Commit
·
a08395c
1
Parent(s):
6c4dc6d
incorporate plotly
Browse files
main.py
CHANGED
|
@@ -1,9 +1,13 @@
|
|
| 1 |
from fasthtml.common import *
|
| 2 |
from fasthtml.components import *
|
| 3 |
from fasthtml.components import D_title, D_article, D_front_matter, D_contents, D_byline
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4 |
|
| 5 |
|
| 6 |
-
app, rt = fast_app(
|
| 7 |
|
| 8 |
|
| 9 |
@app.get("/")
|
|
@@ -12,9 +16,10 @@ def main():
|
|
| 12 |
Head(
|
| 13 |
Meta(charset="UTF-8"),
|
| 14 |
Meta(name="viewport", content="width=device-width, initial-scale=1.0"),
|
| 15 |
-
Link(rel="stylesheet", href="style.css"),
|
| 16 |
Script(src="https://distill.pub/template.v2.js"),
|
| 17 |
Script(src="https://unpkg.com/htmx.org@next/dist/htmx.min.js"),
|
|
|
|
|
|
|
| 18 |
),
|
| 19 |
Body(
|
| 20 |
D_title(
|
|
@@ -76,12 +81,24 @@ def main():
|
|
| 76 |
),
|
| 77 |
Div(
|
| 78 |
A("Web Data", href="#inner-text"),
|
| 79 |
-
hx_get="/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 80 |
hx_target="#inner-text",
|
| 81 |
),
|
| 82 |
-
Div(A("Curated Sources")),
|
| 83 |
-
Div(A("Common Steps")),
|
| 84 |
-
Div(A("TxT360 Results")),
|
| 85 |
role="navigation",
|
| 86 |
cls="l-text figcaption",
|
| 87 |
),
|
|
@@ -235,9 +252,73 @@ def intro():
|
|
| 235 |
)
|
| 236 |
|
| 237 |
|
| 238 |
-
@app.get("/
|
| 239 |
def web_data():
|
| 240 |
return Div(Section(H2(P("Web Data")), id="inner-text"))
|
| 241 |
|
| 242 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 243 |
serve()
|
|
|
|
| 1 |
from fasthtml.common import *
|
| 2 |
from fasthtml.components import *
|
| 3 |
from fasthtml.components import D_title, D_article, D_front_matter, D_contents, D_byline
|
| 4 |
+
from plotly import graph_objects as go
|
| 5 |
+
from fh_plotly import plotly2fasthtml
|
| 6 |
+
import pandas as pd
|
| 7 |
+
import json
|
| 8 |
|
| 9 |
|
| 10 |
+
app, rt = fast_app(debug=True)
|
| 11 |
|
| 12 |
|
| 13 |
@app.get("/")
|
|
|
|
| 16 |
Head(
|
| 17 |
Meta(charset="UTF-8"),
|
| 18 |
Meta(name="viewport", content="width=device-width, initial-scale=1.0"),
|
|
|
|
| 19 |
Script(src="https://distill.pub/template.v2.js"),
|
| 20 |
Script(src="https://unpkg.com/htmx.org@next/dist/htmx.min.js"),
|
| 21 |
+
Script(src="https://cdn.plot.ly/plotly-latest.min.js"),
|
| 22 |
+
Link(rel="stylesheet", href="style.css"),
|
| 23 |
),
|
| 24 |
Body(
|
| 25 |
D_title(
|
|
|
|
| 81 |
),
|
| 82 |
Div(
|
| 83 |
A("Web Data", href="#inner-text"),
|
| 84 |
+
hx_get="/webdata",
|
| 85 |
+
hx_target="#inner-text",
|
| 86 |
+
),
|
| 87 |
+
Div(
|
| 88 |
+
A("Curated Sources", href="#inner-text"),
|
| 89 |
+
hx_get="/curated",
|
| 90 |
+
hx_target="#inner-text",
|
| 91 |
+
),
|
| 92 |
+
Div(
|
| 93 |
+
A("Common Steps", href="#inner-text"),
|
| 94 |
+
hx_get="/common",
|
| 95 |
+
hx_target="#inner-text",
|
| 96 |
+
),
|
| 97 |
+
Div(
|
| 98 |
+
A("TxT360 Results", href="#inner-text"),
|
| 99 |
+
hx_get="/results",
|
| 100 |
hx_target="#inner-text",
|
| 101 |
),
|
|
|
|
|
|
|
|
|
|
| 102 |
role="navigation",
|
| 103 |
cls="l-text figcaption",
|
| 104 |
),
|
|
|
|
| 252 |
)
|
| 253 |
|
| 254 |
|
| 255 |
+
@app.get("/webdata")
|
| 256 |
def web_data():
|
| 257 |
return Div(Section(H2(P("Web Data")), id="inner-text"))
|
| 258 |
|
| 259 |
|
| 260 |
+
def get_chart_28168342():
|
| 261 |
+
fig = go.Figure()
|
| 262 |
+
filter_names = [
|
| 263 |
+
"Download",
|
| 264 |
+
"Language",
|
| 265 |
+
"Min word count",
|
| 266 |
+
"Title Abstract",
|
| 267 |
+
"Majority language",
|
| 268 |
+
"Paragraph count",
|
| 269 |
+
"Frequency",
|
| 270 |
+
"Unigram log probability",
|
| 271 |
+
"Local dedup",
|
| 272 |
+
]
|
| 273 |
+
|
| 274 |
+
data_sources = [
|
| 275 |
+
("Wikipedia", [100, 90, 80, 70, 60, 50, 40, 30, 20]),
|
| 276 |
+
("Freelaw", [100, 90, 80, 70, 60, 50, 40, 20, 20]),
|
| 277 |
+
("DM Maths", [100, 90, 80, 70, 60, 40, 40, 30, 20]),
|
| 278 |
+
("USPTO", [100, 90, 80, 70, 60, 40, 40, 30, 20]),
|
| 279 |
+
("PG19", [100, 90, 80, 70, 60, 40, 40, 30, 20]),
|
| 280 |
+
("Hackernews", [100, 90, 80, 70, 60, 40, 40, 30, 20]),
|
| 281 |
+
("Ubuntu IRC", [100, 90, 80, 70, 60, 40, 40, 30, 20]),
|
| 282 |
+
("Europarl", [100, 90, 80, 70, 60, 40, 40, 30, 20]),
|
| 283 |
+
("StackExchange", [100, 90, 80, 70, 60, 40, 40, 30, 20]),
|
| 284 |
+
("Arxiv", [100, 90, 80, 70, 60, 40, 40, 30, 20]),
|
| 285 |
+
("S2ORC", [100, 90, 80, 70, 60, 40, 40, 30, 20]),
|
| 286 |
+
("S2ORC Abstract", [100, 90, 80, 70, 60, 40, 40, 30, 20]),
|
| 287 |
+
("PubMed Central", [100, 90, 80, 70, 60, 40, 40, 30, 20]),
|
| 288 |
+
("PubMed Central Abstract", [100, 90, 80, 70, 60, 40, 40, 30, 20]),
|
| 289 |
+
("PhilPapers", [100, 90, 80, 70, 60, 40, 40, 30, 20]),
|
| 290 |
+
]
|
| 291 |
+
|
| 292 |
+
for name, x_values in data_sources:
|
| 293 |
+
fig.add_trace(
|
| 294 |
+
go.Funnel(
|
| 295 |
+
name=name,
|
| 296 |
+
orientation="h",
|
| 297 |
+
y=filter_names,
|
| 298 |
+
x=x_values,
|
| 299 |
+
textinfo="value+percent total",
|
| 300 |
+
textposition="inside",
|
| 301 |
+
)
|
| 302 |
+
)
|
| 303 |
+
|
| 304 |
+
# Serialize the Plotly figure to JSON
|
| 305 |
+
fig_json = fig.to_json()
|
| 306 |
+
return fig
|
| 307 |
+
|
| 308 |
+
|
| 309 |
+
@app.get("/curated")
|
| 310 |
+
def curated():
|
| 311 |
+
return Div(plotly2fasthtml(get_chart_28168342()), id="inner-text")
|
| 312 |
+
|
| 313 |
+
|
| 314 |
+
@app.get("/common")
|
| 315 |
+
def common_steps():
|
| 316 |
+
return Div(Section(H2(P("Common Steps")), id="inner-text"))
|
| 317 |
+
|
| 318 |
+
|
| 319 |
+
@app.get("/results")
|
| 320 |
+
def results():
|
| 321 |
+
return Div(Section(H2(P("Results")), id="inner-text"))
|
| 322 |
+
|
| 323 |
+
|
| 324 |
serve()
|
style.css
CHANGED
|
@@ -63,6 +63,7 @@
|
|
| 63 |
gap: 10px;
|
| 64 |
}
|
| 65 |
|
|
|
|
| 66 |
.plotly_slider > input[type="range"] {
|
| 67 |
-webkit-appearance: none;
|
| 68 |
height: 2px;
|
|
@@ -138,6 +139,7 @@ d-byline .byline {
|
|
| 138 |
line-height: 1.8em;
|
| 139 |
}
|
| 140 |
|
|
|
|
| 141 |
@media (min-width: 768px) {
|
| 142 |
d-byline .byline {
|
| 143 |
grid-template-columns: 5fr 1fr 1fr;
|
|
@@ -173,10 +175,11 @@ d-contents a:hover {
|
|
| 173 |
}
|
| 174 |
|
| 175 |
|
| 176 |
-
@media (min-width:
|
| 177 |
d-article {
|
| 178 |
/* Ensure d-article does not prevent sticky positioning */
|
| 179 |
-
overflow-x: visible
|
|
|
|
| 180 |
}
|
| 181 |
|
| 182 |
d-contents {
|
|
|
|
| 63 |
gap: 10px;
|
| 64 |
}
|
| 65 |
|
| 66 |
+
|
| 67 |
.plotly_slider > input[type="range"] {
|
| 68 |
-webkit-appearance: none;
|
| 69 |
height: 2px;
|
|
|
|
| 139 |
line-height: 1.8em;
|
| 140 |
}
|
| 141 |
|
| 142 |
+
|
| 143 |
@media (min-width: 768px) {
|
| 144 |
d-byline .byline {
|
| 145 |
grid-template-columns: 5fr 1fr 1fr;
|
|
|
|
| 175 |
}
|
| 176 |
|
| 177 |
|
| 178 |
+
@media (min-width: 1100px) {
|
| 179 |
d-article {
|
| 180 |
/* Ensure d-article does not prevent sticky positioning */
|
| 181 |
+
overflow-x: visible;
|
| 182 |
+
grid-template-columns: [page-start] 1fr [page-start kicker-start] 1fr [middle-start] 1fr [text-start kicker-end] 1fr 1fr 1fr 1fr 1fr 1fr 1fr 1fr 1fr 1fr [text-end gutter-start] 1fr [middle-end] 0fr [page-end gutter-end] 1fr [screen-end];
|
| 183 |
}
|
| 184 |
|
| 185 |
d-contents {
|