feat: plots update + usage%embodied
Browse files- app.py +27 -43
- src/utils.py +17 -2
app.py
CHANGED
@@ -5,9 +5,6 @@ from bs4 import BeautifulSoup
|
|
5 |
|
6 |
import tiktoken
|
7 |
|
8 |
-
import matplotlib
|
9 |
-
import matplotlib.pyplot as plt
|
10 |
-
|
11 |
from ecologits.tracers.utils import compute_llm_impacts, _avg
|
12 |
from ecologits.impacts.llm import compute_llm_impacts as compute_llm_impacts_expert
|
13 |
from ecologits.impacts.llm import IF_ELECTRICITY_MIX_GWP, IF_ELECTRICITY_MIX_ADPE, IF_ELECTRICITY_MIX_PE
|
@@ -35,11 +32,13 @@ from src.constants import (
|
|
35 |
)
|
36 |
from src.utils import (
|
37 |
format_impacts,
|
|
|
38 |
format_energy_eq_physical_activity,
|
39 |
PhysicalActivity,
|
40 |
format_energy_eq_electric_vehicle,
|
41 |
format_gwp_eq_streaming, format_energy_eq_electricity_production, EnergyProduction,
|
42 |
-
format_gwp_eq_airplane_paris_nyc, format_energy_eq_electricity_consumption_ireland
|
|
|
43 |
)
|
44 |
|
45 |
CUSTOM = "Custom"
|
@@ -364,10 +363,10 @@ with gr.Blocks(css=custom_css) as demo:
|
|
364 |
if_electricity_mix_adpe=mix_adpe,
|
365 |
if_electricity_mix_pe=mix_pe
|
366 |
)
|
367 |
-
impacts =
|
368 |
|
369 |
with gr.Blocks():
|
370 |
-
|
371 |
with gr.Row():
|
372 |
gr.Markdown(f"""
|
373 |
<h2 align = "center">Environmental impacts</h2>
|
@@ -380,61 +379,46 @@ with gr.Blocks(css=custom_css) as demo:
|
|
380 |
$$ \Large {impacts.energy.magnitude:.3g} \ \large {impacts.energy.units} $$
|
381 |
<p align="center"><i>Evaluates the electricity consumption<i></p><br>
|
382 |
""")
|
|
|
383 |
with gr.Column(scale=1, min_width=220):
|
384 |
gr.Markdown(f"""
|
385 |
<h2 align="center">🌍️ GHG Emissions</h2>
|
386 |
$$ \Large {impacts.gwp.magnitude:.3g} \ \large {impacts.gwp.units} $$
|
387 |
<p align="center"><i>Evaluates the effect on global warming<i></p><br>
|
|
|
|
|
388 |
""")
|
|
|
389 |
with gr.Column(scale=1, min_width=220):
|
390 |
gr.Markdown(f"""
|
391 |
<h2 align="center">🪨 Abiotic Resources</h2>
|
392 |
$$ \Large {impacts.adpe.magnitude:.3g} \ \large {impacts.adpe.units} $$
|
393 |
<p align="center"><i>Evaluates the use of metals and minerals<i></p><br>
|
|
|
|
|
394 |
""")
|
|
|
395 |
with gr.Column(scale=1, min_width=220):
|
396 |
gr.Markdown(f"""
|
397 |
<h2 align="center">⛽️ Primary Energy</h2>
|
398 |
$$ \Large {impacts.pe.magnitude:.3g} \ \large {impacts.pe.units} $$
|
399 |
<p align="center"><i>Evaluates the use of energy resources<i></p><br>
|
|
|
|
|
400 |
""")
|
401 |
-
|
402 |
-
|
403 |
-
|
404 |
-
|
405 |
-
|
406 |
-
|
407 |
-
|
408 |
-
|
409 |
-
|
410 |
-
|
411 |
-
|
412 |
-
|
413 |
-
|
414 |
-
plt.text(i, y[i], y[i], ha = 'center')
|
415 |
-
|
416 |
-
fig, ax = plt.subplots(figsize=(15,5), facecolor='#1F2937')
|
417 |
-
ax.bar(categories, values)
|
418 |
-
#ax.set_xlabel('Countries')
|
419 |
-
ax.set_ylabel('GHG emissions (gCO2eq) for 1kWh')
|
420 |
-
ax.set_title('GWP emissions for 1 kWh of electricity consumption')
|
421 |
-
ax.set_facecolor("#0B0F19")
|
422 |
-
|
423 |
-
addlabels(categories, values)
|
424 |
-
|
425 |
-
font = {'family' : 'monospace',
|
426 |
-
'weight' : 'normal',
|
427 |
-
'size' : 14}
|
428 |
-
|
429 |
-
matplotlib.rc('font', **font)
|
430 |
-
matplotlib.rcParams.update({'text.color':'white',
|
431 |
-
'axes.labelcolor':'white',
|
432 |
-
'xtick.color':'white',
|
433 |
-
'ytick.color':'white'})
|
434 |
-
|
435 |
-
return fig
|
436 |
-
|
437 |
-
static_plot = gr.Plot(value=create_static_bar_plot())
|
438 |
|
439 |
with gr.Tab("🔍 Evaluate your own usage"):
|
440 |
|
|
|
5 |
|
6 |
import tiktoken
|
7 |
|
|
|
|
|
|
|
8 |
from ecologits.tracers.utils import compute_llm_impacts, _avg
|
9 |
from ecologits.impacts.llm import compute_llm_impacts as compute_llm_impacts_expert
|
10 |
from ecologits.impacts.llm import IF_ELECTRICITY_MIX_GWP, IF_ELECTRICITY_MIX_ADPE, IF_ELECTRICITY_MIX_PE
|
|
|
32 |
)
|
33 |
from src.utils import (
|
34 |
format_impacts,
|
35 |
+
format_impacts_expert,
|
36 |
format_energy_eq_physical_activity,
|
37 |
PhysicalActivity,
|
38 |
format_energy_eq_electric_vehicle,
|
39 |
format_gwp_eq_streaming, format_energy_eq_electricity_production, EnergyProduction,
|
40 |
+
format_gwp_eq_airplane_paris_nyc, format_energy_eq_electricity_consumption_ireland,
|
41 |
+
df_elec_mix_for_plot
|
42 |
)
|
43 |
|
44 |
CUSTOM = "Custom"
|
|
|
363 |
if_electricity_mix_adpe=mix_adpe,
|
364 |
if_electricity_mix_pe=mix_pe
|
365 |
)
|
366 |
+
impacts, usage, embodied = format_impacts_expert(impacts)
|
367 |
|
368 |
with gr.Blocks():
|
369 |
+
|
370 |
with gr.Row():
|
371 |
gr.Markdown(f"""
|
372 |
<h2 align = "center">Environmental impacts</h2>
|
|
|
379 |
$$ \Large {impacts.energy.magnitude:.3g} \ \large {impacts.energy.units} $$
|
380 |
<p align="center"><i>Evaluates the electricity consumption<i></p><br>
|
381 |
""")
|
382 |
+
|
383 |
with gr.Column(scale=1, min_width=220):
|
384 |
gr.Markdown(f"""
|
385 |
<h2 align="center">🌍️ GHG Emissions</h2>
|
386 |
$$ \Large {impacts.gwp.magnitude:.3g} \ \large {impacts.gwp.units} $$
|
387 |
<p align="center"><i>Evaluates the effect on global warming<i></p><br>
|
388 |
+
$$ \Large {100*usage.gwp.value / (usage.gwp.value + embodied.gwp.value):.3} $$
|
389 |
+
<p align="center"><i>% of GWP by usage (vs embodied)<i></p><br>
|
390 |
""")
|
391 |
+
|
392 |
with gr.Column(scale=1, min_width=220):
|
393 |
gr.Markdown(f"""
|
394 |
<h2 align="center">🪨 Abiotic Resources</h2>
|
395 |
$$ \Large {impacts.adpe.magnitude:.3g} \ \large {impacts.adpe.units} $$
|
396 |
<p align="center"><i>Evaluates the use of metals and minerals<i></p><br>
|
397 |
+
$$ \Large {100*usage.adpe.value / (usage.adpe.value + embodied.adpe.value):.3} $$
|
398 |
+
<p align="center"><i>% of ADPE by usage (vs embodied)<i></p><br>
|
399 |
""")
|
400 |
+
|
401 |
with gr.Column(scale=1, min_width=220):
|
402 |
gr.Markdown(f"""
|
403 |
<h2 align="center">⛽️ Primary Energy</h2>
|
404 |
$$ \Large {impacts.pe.magnitude:.3g} \ \large {impacts.pe.units} $$
|
405 |
<p align="center"><i>Evaluates the use of energy resources<i></p><br>
|
406 |
+
$$ \Large {100*usage.pe.value / (usage.pe.value + embodied.pe.value):.3} $$
|
407 |
+
<p align="center"><i>% of PE by usage (vs embodied)<i></p><br>
|
408 |
""")
|
409 |
+
|
410 |
+
with gr.Row():
|
411 |
+
gr.Markdown(f"""
|
412 |
+
<h2 align="center">How can location impact the footprint ?</h2>
|
413 |
+
""")
|
414 |
+
|
415 |
+
with gr.Row():
|
416 |
+
gr.BarPlot(df_elec_mix_for_plot,
|
417 |
+
x='country',
|
418 |
+
y='electricity_mix',
|
419 |
+
sort='y',
|
420 |
+
x_title=None,
|
421 |
+
y_title='electricity mix in gCO2eq / kWh')
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
422 |
|
423 |
with gr.Tab("🔍 Evaluate your own usage"):
|
424 |
|
src/utils.py
CHANGED
@@ -1,6 +1,8 @@
|
|
1 |
from dataclasses import dataclass
|
2 |
from enum import Enum
|
3 |
|
|
|
|
|
4 |
from ecologits.impacts.models import Impacts, Energy, GWP, ADPe, PE
|
5 |
from pint import UnitRegistry, Quantity
|
6 |
|
@@ -50,6 +52,11 @@ COUNTRIES = [
|
|
50 |
("samoa", 100, 821_632),
|
51 |
]
|
52 |
|
|
|
|
|
|
|
|
|
|
|
53 |
|
54 |
# From https://www.runningtools.com/energyusage.htm
|
55 |
RUNNING_ENERGY_EQ = q("294 kJ / km") # running 1 km at 10 km/h with a weight of 70 kg
|
@@ -108,15 +115,23 @@ def format_pe(pe: PE) -> Quantity:
|
|
108 |
val = val.to("kJ")
|
109 |
return val
|
110 |
|
111 |
-
|
112 |
def format_impacts(impacts: Impacts) -> QImpacts:
|
113 |
return QImpacts(
|
114 |
energy=format_energy(impacts.energy),
|
115 |
gwp=format_gwp(impacts.gwp),
|
116 |
adpe=format_adpe(impacts.adpe),
|
117 |
-
pe=format_pe(impacts.pe)
|
118 |
)
|
119 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
120 |
|
121 |
def format_energy_eq_physical_activity(energy: Quantity) -> tuple[PhysicalActivity, Quantity]:
|
122 |
energy = energy.to("kJ")
|
|
|
1 |
from dataclasses import dataclass
|
2 |
from enum import Enum
|
3 |
|
4 |
+
import pandas as pd
|
5 |
+
|
6 |
from ecologits.impacts.models import Impacts, Energy, GWP, ADPe, PE
|
7 |
from pint import UnitRegistry, Quantity
|
8 |
|
|
|
52 |
("samoa", 100, 821_632),
|
53 |
]
|
54 |
|
55 |
+
def df_elec_mix_for_plot():
|
56 |
+
return pd.DataFrame({
|
57 |
+
'country': ['Sweden', 'France', 'Canada', 'USA', 'China', 'Australia', 'India'],
|
58 |
+
'electricity_mix': [46, 81, 238, 679, 1057, 1123, 1583]
|
59 |
+
})
|
60 |
|
61 |
# From https://www.runningtools.com/energyusage.htm
|
62 |
RUNNING_ENERGY_EQ = q("294 kJ / km") # running 1 km at 10 km/h with a weight of 70 kg
|
|
|
115 |
val = val.to("kJ")
|
116 |
return val
|
117 |
|
|
|
118 |
def format_impacts(impacts: Impacts) -> QImpacts:
|
119 |
return QImpacts(
|
120 |
energy=format_energy(impacts.energy),
|
121 |
gwp=format_gwp(impacts.gwp),
|
122 |
adpe=format_adpe(impacts.adpe),
|
123 |
+
pe=format_pe(impacts.pe)
|
124 |
)
|
125 |
|
126 |
+
def format_impacts_expert(impacts: Impacts) -> QImpacts:
|
127 |
+
return QImpacts(
|
128 |
+
energy=format_energy(impacts.energy),
|
129 |
+
gwp=format_gwp(impacts.gwp),
|
130 |
+
adpe=format_adpe(impacts.adpe),
|
131 |
+
pe=format_pe(impacts.pe)
|
132 |
+
|
133 |
+
), impacts.usage, impacts.embodied
|
134 |
+
|
135 |
|
136 |
def format_energy_eq_physical_activity(energy: Quantity) -> tuple[PhysicalActivity, Quantity]:
|
137 |
energy = energy.to("kJ")
|