go
Browse files- app.py +39 -30
- layouts/app.grid.json +119 -0
app.py
CHANGED
|
@@ -1,5 +1,5 @@
|
|
| 1 |
# /// script
|
| 2 |
-
# requires-python = "==3.
|
| 3 |
# dependencies = [
|
| 4 |
# "marimo",
|
| 5 |
# "polars==1.23.0",
|
|
@@ -16,7 +16,7 @@
|
|
| 16 |
import marimo
|
| 17 |
|
| 18 |
__generated_with = "0.11.9"
|
| 19 |
-
app = marimo.App(width="medium")
|
| 20 |
|
| 21 |
|
| 22 |
@app.cell
|
|
@@ -27,9 +27,7 @@ def _(mo):
|
|
| 27 |
|
| 28 |
@app.cell
|
| 29 |
def _(mo, use_default_switch):
|
| 30 |
-
mo.
|
| 31 |
-
|
| 32 |
-
uploaded_file = mo.ui.file(kind="area")
|
| 33 |
uploaded_file
|
| 34 |
return (uploaded_file,)
|
| 35 |
|
|
@@ -49,14 +47,20 @@ def _(mo):
|
|
| 49 |
|
| 50 |
|
| 51 |
@app.cell
|
| 52 |
-
def _(
|
| 53 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 54 |
|
| 55 |
if use_default_switch.value:
|
| 56 |
df = pl.read_csv("spam.csv")
|
| 57 |
else:
|
| 58 |
-
df = pl.read_csv(
|
| 59 |
-
|
| 60 |
texts = df["text"].to_list()
|
| 61 |
return df, texts
|
| 62 |
|
|
@@ -151,12 +155,6 @@ def _():
|
|
| 151 |
return (br,)
|
| 152 |
|
| 153 |
|
| 154 |
-
@app.cell
|
| 155 |
-
def _(get_label):
|
| 156 |
-
get_label()
|
| 157 |
-
return
|
| 158 |
-
|
| 159 |
-
|
| 160 |
@app.cell
|
| 161 |
def _(mo, neg_label, pos_label):
|
| 162 |
get_label, set_label = mo.state({pos_label.value: [], neg_label.value: []})
|
|
@@ -205,6 +203,7 @@ def _(df_emb, mo, scatter):
|
|
| 205 |
@app.cell
|
| 206 |
def _(mo):
|
| 207 |
switch = mo.ui.switch(False, label="Use search")
|
|
|
|
| 208 |
return (switch,)
|
| 209 |
|
| 210 |
|
|
@@ -268,24 +267,34 @@ def _(
|
|
| 268 |
return df_emb, ham, i, labels, query, similarity, spam
|
| 269 |
|
| 270 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 271 |
@app.cell
|
| 272 |
def _():
|
| 273 |
import marimo as mo
|
| 274 |
-
|
| 275 |
-
|
| 276 |
-
|
| 277 |
-
|
| 278 |
-
|
| 279 |
-
|
| 280 |
-
return (
|
| 281 |
-
LogisticRegression,
|
| 282 |
-
SentenceTransformer,
|
| 283 |
-
alt,
|
| 284 |
-
cosine_similarity,
|
| 285 |
-
mo,
|
| 286 |
-
np,
|
| 287 |
-
pl,
|
| 288 |
-
)
|
| 289 |
|
| 290 |
|
| 291 |
if __name__ == "__main__":
|
|
|
|
| 1 |
# /// script
|
| 2 |
+
# requires-python = "==3.12"
|
| 3 |
# dependencies = [
|
| 4 |
# "marimo",
|
| 5 |
# "polars==1.23.0",
|
|
|
|
| 16 |
import marimo
|
| 17 |
|
| 18 |
__generated_with = "0.11.9"
|
| 19 |
+
app = marimo.App(width="medium", layout_file="layouts/app.grid.json")
|
| 20 |
|
| 21 |
|
| 22 |
@app.cell
|
|
|
|
| 27 |
|
| 28 |
@app.cell
|
| 29 |
def _(mo, use_default_switch):
|
| 30 |
+
uploaded_file = mo.ui.file(kind="area") if not use_default_switch.value else None
|
|
|
|
|
|
|
| 31 |
uploaded_file
|
| 32 |
return (uploaded_file,)
|
| 33 |
|
|
|
|
| 47 |
|
| 48 |
|
| 49 |
@app.cell
|
| 50 |
+
def _(uploaded_file, use_default_switch):
|
| 51 |
+
should_stop = not use_default_switch.value and len(uploaded_file.value) == 0
|
| 52 |
+
return (should_stop,)
|
| 53 |
+
|
| 54 |
+
|
| 55 |
+
@app.cell
|
| 56 |
+
def _(mo, pl, should_stop, uploaded_file, use_default_switch):
|
| 57 |
+
mo.stop(should_stop , mo.md("**Submit a dataset or use default one to continue.**"))
|
| 58 |
|
| 59 |
if use_default_switch.value:
|
| 60 |
df = pl.read_csv("spam.csv")
|
| 61 |
else:
|
| 62 |
+
df = pl.read_csv(uploaded_file.value[0].contents)
|
| 63 |
+
|
| 64 |
texts = df["text"].to_list()
|
| 65 |
return df, texts
|
| 66 |
|
|
|
|
| 155 |
return (br,)
|
| 156 |
|
| 157 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 158 |
@app.cell
|
| 159 |
def _(mo, neg_label, pos_label):
|
| 160 |
get_label, set_label = mo.state({pos_label.value: [], neg_label.value: []})
|
|
|
|
| 203 |
@app.cell
|
| 204 |
def _(mo):
|
| 205 |
switch = mo.ui.switch(False, label="Use search")
|
| 206 |
+
switch
|
| 207 |
return (switch,)
|
| 208 |
|
| 209 |
|
|
|
|
| 267 |
return df_emb, ham, i, labels, query, similarity, spam
|
| 268 |
|
| 269 |
|
| 270 |
+
@app.cell
|
| 271 |
+
def _(mo):
|
| 272 |
+
|
| 273 |
+
with mo.status.spinner(subtitle="Loading libraries ...") as _spinner:
|
| 274 |
+
import polars as pl
|
| 275 |
+
import altair as alt
|
| 276 |
+
import numpy as np
|
| 277 |
+
from sklearn.metrics.pairwise import cosine_similarity
|
| 278 |
+
from sklearn.linear_model import LogisticRegression
|
| 279 |
+
return LogisticRegression, alt, cosine_similarity, np, pl
|
| 280 |
+
|
| 281 |
+
|
| 282 |
+
@app.cell
|
| 283 |
+
def _(mo):
|
| 284 |
+
with mo.status.spinner(subtitle="Loading SBERT ...") as _spinner:
|
| 285 |
+
from sentence_transformers import SentenceTransformer
|
| 286 |
+
return (SentenceTransformer,)
|
| 287 |
+
|
| 288 |
+
|
| 289 |
@app.cell
|
| 290 |
def _():
|
| 291 |
import marimo as mo
|
| 292 |
+
return (mo,)
|
| 293 |
+
|
| 294 |
+
|
| 295 |
+
@app.cell
|
| 296 |
+
def _():
|
| 297 |
+
return
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 298 |
|
| 299 |
|
| 300 |
if __name__ == "__main__":
|
layouts/app.grid.json
ADDED
|
@@ -0,0 +1,119 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"type": "grid",
|
| 3 |
+
"data": {
|
| 4 |
+
"columns": 24,
|
| 5 |
+
"rowHeight": 20,
|
| 6 |
+
"maxWidth": 1400,
|
| 7 |
+
"bordered": true,
|
| 8 |
+
"cells": [
|
| 9 |
+
{
|
| 10 |
+
"position": [
|
| 11 |
+
0,
|
| 12 |
+
0,
|
| 13 |
+
6,
|
| 14 |
+
2
|
| 15 |
+
]
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"position": [
|
| 19 |
+
6,
|
| 20 |
+
0,
|
| 21 |
+
11,
|
| 22 |
+
8
|
| 23 |
+
]
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"position": [
|
| 27 |
+
0,
|
| 28 |
+
5,
|
| 29 |
+
6,
|
| 30 |
+
3
|
| 31 |
+
]
|
| 32 |
+
},
|
| 33 |
+
{
|
| 34 |
+
"position": null
|
| 35 |
+
},
|
| 36 |
+
{
|
| 37 |
+
"position": null
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"position": null
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"position": null
|
| 44 |
+
},
|
| 45 |
+
{
|
| 46 |
+
"position": null
|
| 47 |
+
},
|
| 48 |
+
{
|
| 49 |
+
"position": null
|
| 50 |
+
},
|
| 51 |
+
{
|
| 52 |
+
"position": null
|
| 53 |
+
},
|
| 54 |
+
{
|
| 55 |
+
"position": [
|
| 56 |
+
0,
|
| 57 |
+
8,
|
| 58 |
+
13,
|
| 59 |
+
37
|
| 60 |
+
]
|
| 61 |
+
},
|
| 62 |
+
{
|
| 63 |
+
"position": [
|
| 64 |
+
13,
|
| 65 |
+
18,
|
| 66 |
+
11,
|
| 67 |
+
23
|
| 68 |
+
]
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"position": null
|
| 72 |
+
},
|
| 73 |
+
{
|
| 74 |
+
"position": null
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"position": null
|
| 78 |
+
},
|
| 79 |
+
{
|
| 80 |
+
"position": null
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"position": [
|
| 84 |
+
13,
|
| 85 |
+
16,
|
| 86 |
+
11,
|
| 87 |
+
2
|
| 88 |
+
]
|
| 89 |
+
},
|
| 90 |
+
{
|
| 91 |
+
"position": null
|
| 92 |
+
},
|
| 93 |
+
{
|
| 94 |
+
"position": null
|
| 95 |
+
},
|
| 96 |
+
{
|
| 97 |
+
"position": null
|
| 98 |
+
},
|
| 99 |
+
{
|
| 100 |
+
"position": null
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"position": null
|
| 104 |
+
},
|
| 105 |
+
{
|
| 106 |
+
"position": null
|
| 107 |
+
},
|
| 108 |
+
{
|
| 109 |
+
"position": null
|
| 110 |
+
},
|
| 111 |
+
{
|
| 112 |
+
"position": null
|
| 113 |
+
},
|
| 114 |
+
{
|
| 115 |
+
"position": null
|
| 116 |
+
}
|
| 117 |
+
]
|
| 118 |
+
}
|
| 119 |
+
}
|