if metadata['takerToken'] != 'ETH':
continue
fees = 0
if args.fees:
# Can't differentiate fees from assets in the following scenarios.
if metadata['makerToken'] == 'ETH':
continue
if metadata['side'] == 'buy' and metadata['takerToken'] == 'ETH':
continue
# Express fees in units of gas
fees = int(swap['protocolFee']) - int(
metadata['swapResult']['ethBalance'])
fees = max(0, fees // int(swap['gasPrice']))
costs[url] = gas = metadata['swapResult']['gasUsed'] + fees
if len(urls) == len(list(costs.keys())):
value = get_max_value(swap)
for url in costs:
costs_by_url_by_value[url] = costs_by_url_by_value.get(url, {})
costs_by_url_by_value[url][value] = costs_by_url_by_value[url].get(
value, [])
costs_by_url_by_value[url][value].append(costs[url])
sns.catplot(
x='value',
y='gas used',
hue='url',
data=pd.DataFrame(
[[
url, max_value,
sum(costs_by_url_by_value[url].get(max_value, [])) /
(len(costs_by_url_by_value[url].get(max_value, [])) or 1)
for d in data]
], )))
print(f'Found {len(sources)} sources')
data_by_source = {
s: [
d for d in data
if len(d['sources']) == 1 and d['sources'][0]['name'] == s
]
for s in sources
}
df_values = []
for s in sources:
for d in data_by_source[s]:
df_values.append([s, get_max_value(d), float(get_slippage(d))])
tokens = sorted(
set([
*(d['metadata']['makerToken'] for d in data),
*(d['metadata']['takerToken'] for d in data)
]))
sns.catplot(
x='source',
y='slippage',
hue='swap value',
data=pd.DataFrame(
df_values,
columns=['source', 'swap value', 'slippage'],
),
kind='bar',
args = get_program_args()
data = [d for d in load_ab_data(args.path) if len(list(d.keys())) > 1]
print(f'Loaded {len(data)} data items')
urls = sorted(set(itertools.chain(*(d.keys() for d in data))))
sns.catplot(
x='swap value',
y='accuracy',
hue='url',
data=pd.DataFrame(
[[
url,
float(d[url]['metadata']['swapResult']['boughtAmount']) /
float(d[url]['buyAmount']) if d.get(url) else None,
get_max_value(d[url]) if d.get(url) else None,
] for url, d in itertools.product(urls, data)],
columns=['url', 'accuracy', 'swap value'],
),
kind='bar',
errcolor='black',
errwidth=1,
capsize=.1,
order=[max_value for min_value, max_value in VALUES],
legend=True,
legend_out=False,
)
for t, url in zip(plt.gca().get_legend().texts, urls):
t.set_text(url)
plt.xticks(list(range(len(VALUES))),
print(f'Loaded {len(data)} data items')
tokens = sorted(
set([
*(d['metadata']['makerToken'] for d in data),
*(d['metadata']['takerToken'] for d in data)
]))
sns.catplot(
x='token',
y='slippage',
hue='swap value',
data=pd.DataFrame(
[
*([
d['metadata']['takerToken'],
get_max_value(d),
float(get_slippage(d)),
] for d in data), *([
d['metadata']['makerToken'],
get_max_value(d),
float(get_slippage(d)),
] for d in data)
],
columns=['token', 'swap value', 'slippage'],
),
kind='bar',
errcolor='black',
errwidth=1,
capsize=.1,
order=tokens,
legend=True,
return True
args = get_program_args()
data = [
d for d in load_ab_data(args.path) if are_valid_swaps(args, d.values())
]
print(f'Loaded {len(data)} data items')
count_by_value_by_url = {}
for d in data:
best_swap_url = max(d.keys(), key=lambda k: get_realized_price(d[k]))
worst_swap_url = min(d.keys(), key=lambda k: get_realized_price(d[k]))
if best_swap_url == worst_swap_url:
continue
best_swap = d[best_swap_url]
worst_swap = d[worst_swap_url]
best_price = get_realized_price(best_swap)
worst_price = get_realized_price(worst_swap)
bps = (best_price - worst_price) / worst_price * 1e4
value = get_max_value(d[best_swap_url])
count_by_value_by_url[value] = count_by_value_by_url.get(value, {})
count_by_value_by_url[value][best_swap_url] = count_by_value_by_url[
value].get(best_swap_url, 0)
count_by_value_by_url[value][worst_swap_url] = count_by_value_by_url[
value].get(worst_swap_url, 0)
count_by_value_by_url[value][best_swap_url] += 1
print(count_by_value_by_url)
args = get_program_args()
data = [d for d in load_ab_data(args.path) if len(list(d.keys())) > 1]
print(f'Loaded {len(data)} data items')
urls = sorted(set(itertools.chain(*(d.keys() for d in data))))
sns.catplot(
x='swap value',
y='response time',
hue='url',
data=pd.DataFrame(
[[
url,
d[url]['metadata']['responseTime'],
get_max_value(d[url]),
] for url, d in itertools.product(urls, data)],
columns=['url', 'response time', 'swap value'],
),
kind='bar',
errcolor='black',
errwidth=1,
capsize=.1,
order=[max_value for min_value, max_value in VALUES],
legend=True,
legend_out=False,
)
for t, url in zip(plt.gca().get_legend().texts, urls):
t.set_text(url)
plt.xticks(list(range(len(VALUES))),
