-
Notifications
You must be signed in to change notification settings - Fork 747
/
screener.py
executable file
·421 lines (387 loc) · 19.8 KB
/
screener.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
import time
import json
import pandas as pd
import re
import sys
from datetime import datetime
from decimal import Decimal
from itertools import islice
from tradingview_ta import *
from importlib.metadata import version
from controllers.PyCryptoBot import PyCryptoBot
from models.helper.TelegramBotHelper import TelegramBotHelper as TGBot
from models.exchange.binance import PublicAPI as BPublicAPI
from models.exchange.coinbase import AuthAPI as CBAuthAPI
from models.exchange.coinbase_pro import PublicAPI as CPublicAPI
from models.exchange.kucoin import PublicAPI as KPublicAPI
from models.exchange.Granularity import Granularity
from models.exchange.ExchangesEnum import Exchange as CryptoExchange
def volatility_calculator(bollinger_band_upper, bollinger_band_lower, keltner_upper, keltner_lower, high, low):
"""
A break away from traditional volatility calculations. Based entirely
on the proportionate self.price gap between keltner channels, bolinger, and high / low averaged out
"""
try:
b_spread = Decimal(bollinger_band_upper) - Decimal(bollinger_band_lower)
k_spread = Decimal(keltner_upper) - Decimal(keltner_lower)
p_spread = Decimal(high) - Decimal(low)
except TypeError:
return 0
b_pcnt = abs(b_spread / Decimal(bollinger_band_lower)) * 100
k_pcnt = abs(k_spread / Decimal(keltner_lower)) * 100
chan_20_pcnt = (b_pcnt + k_pcnt) / 2
p_pcnt = abs(p_spread / Decimal(low)) * 100
return abs((chan_20_pcnt + p_pcnt) / 2)
def load_configs():
exchanges_loaded = []
try:
with open("screener.json", encoding="utf8") as json_file:
config = json.load(json_file)
except IOError as err:
raise (err)
try:
with open("config.json", encoding="utf8") as json_file:
bot_config = json.load(json_file)
except IOError as err:
print(err)
try:
for exchange in config:
ex = CryptoExchange(exchange)
exchange_config = config[ex.value]
if ex == CryptoExchange.BINANCE:
binance_app = PyCryptoBot(exchange=ex)
binance_app.public_api = BPublicAPI(bot_config[ex.value]["api_url"])
binance_app.scanner_quote_currencies = exchange_config.get("quote_currency", ["USDT"])
binance_app.granularity = Granularity(Granularity.convert_to_enum(exchange_config.get("granularity", "1h")))
binance_app.adx_threshold = exchange_config.get("adx_threshold", 25)
binance_app.volatility_threshold = exchange_config.get("volatility_threshold", 9)
binance_app.minimum_volatility = exchange_config.get("minimum_volatility", 5)
binance_app.minimum_volume = exchange_config.get("minimum_volume", 20000)
binance_app.volume_threshold = exchange_config.get("volume_threshold", 20000)
binance_app.minimum_quote_price = exchange_config.get("minimum_quote_price", 0.0000001)
binance_app.selection_score = exchange_config.get("selection_score", 10)
binance_app.tv_screener_ratings = [rating.upper() for rating in exchange_config.get("tv_screener_ratings", ["STRONG_BUY"])]
exchanges_loaded.append(binance_app)
elif ex == CryptoExchange.COINBASE:
coinbase_app = PyCryptoBot(exchange=ex)
coinbase_app.public_api = CBAuthAPI(bot_config[ex.value]["api_key"], bot_config[ex.value]["api_secret"], bot_config[ex.value]["api_url"])
coinbase_app.scanner_quote_currencies = exchange_config.get("quote_currency", ["USDT"])
coinbase_app.granularity = Granularity(Granularity.convert_to_enum(int(exchange_config.get("granularity", "3600"))))
coinbase_app.adx_threshold = exchange_config.get("adx_threshold", 25)
coinbase_app.volatility_threshold = exchange_config.get("volatility_threshold", 9)
coinbase_app.minimum_volatility = exchange_config.get("minimum_volatility", 5)
coinbase_app.minimum_volume = exchange_config.get("minimum_volume", 20000)
coinbase_app.volume_threshold = exchange_config.get("volume_threshold", 20000)
coinbase_app.minimum_quote_price = exchange_config.get("minimum_quote_price", 0.0000001)
coinbase_app.selection_score = exchange_config.get("selection_score", 10)
coinbase_app.tv_screener_ratings = [rating.upper() for rating in exchange_config.get("tv_screener_ratings", ["STRONG_BUY"])]
exchanges_loaded.append(coinbase_app)
elif ex == CryptoExchange.COINBASEPRO:
coinbase_pro_app = PyCryptoBot(exchange=ex)
coinbase_pro_app.public_api = CPublicAPI()
coinbase_pro_app.scanner_quote_currencies = exchange_config.get("quote_currency", ["USDT"])
coinbase_pro_app.granularity = Granularity(Granularity.convert_to_enum(int(exchange_config.get("granularity", "3600"))))
coinbase_pro_app.adx_threshold = exchange_config.get("adx_threshold", 25)
coinbase_pro_app.volatility_threshold = exchange_config.get("volatility_threshold", 9)
coinbase_pro_app.minimum_volatility = exchange_config.get("minimum_volatility", 5)
coinbase_pro_app.minimum_volume = exchange_config.get("minimum_volume", 20000)
coinbase_pro_app.volume_threshold = exchange_config.get("volume_threshold", 20000)
coinbase_pro_app.minimum_quote_price = exchange_config.get("minimum_quote_price", 0.0000001)
coinbase_pro_app.selection_score = exchange_config.get("selection_score", 10)
coinbase_pro_app.tv_screener_ratings = [rating.upper() for rating in exchange_config.get("tv_screener_ratings", ["STRONG_BUY"])]
exchanges_loaded.append(coinbase_pro_app)
elif ex == CryptoExchange.KUCOIN:
kucoin_app = PyCryptoBot(exchange=ex)
kucoin_app.public_api = KPublicAPI(bot_config[ex.value]["api_url"])
kucoin_app.scanner_quote_currencies = exchange_config.get("quote_currency", ["USDT"])
kucoin_app.granularity = Granularity(Granularity.convert_to_enum(exchange_config.get("granularity", "1h")))
kucoin_app.adx_threshold = exchange_config.get("adx_threshold", 25)
kucoin_app.volatility_threshold = exchange_config.get("volatility_threshold", 9)
kucoin_app.minimum_volatility = exchange_config.get("minimum_volatility", 5)
kucoin_app.minimum_volume = exchange_config.get("minimum_volume", 20000)
kucoin_app.volume_threshold = exchange_config.get("volume_threshold", 20000)
kucoin_app.minimum_quote_price = exchange_config.get("minimum_quote_price", 0.0000001)
kucoin_app.selection_score = exchange_config.get("selection_score", 10)
kucoin_app.tv_screener_ratings = [rating.upper() for rating in exchange_config.get("tv_screener_ratings", ["STRONG_BUY"])]
exchanges_loaded.append(kucoin_app)
else:
raise ValueError(f"Invalid exchange found in config: {ex}")
except AttributeError as e:
print(f"Invalid exchange: {e}...ignoring.")
return exchanges_loaded
def chunker(market_list, chunk_size):
markets = iter(market_list)
market_chunk = list(islice(markets, chunk_size))
while market_chunk:
yield market_chunk
market_chunk = list(islice(markets, chunk_size))
def get_markets(app, quote_currency):
markets = []
quote_currency = quote_currency.upper()
api = app.public_api
resp = api.get_markets_24hr_stats()
if app.exchange == CryptoExchange.BINANCE:
for row in resp:
if row["symbol"].endswith(quote_currency):
markets.append(row["symbol"])
elif app.exchange == CryptoExchange.COINBASE:
for market in resp:
market = str(market)
if market.endswith(f"-{quote_currency}"):
markets.append(market)
elif app.exchange == CryptoExchange.COINBASEPRO:
for market in resp:
market = str(market)
if market.endswith(f"-{quote_currency}"):
markets.append(market)
elif app.exchange == CryptoExchange.KUCOIN:
results = resp["data"]["ticker"]
for result in results:
if result["symbol"].endswith(f"-{quote_currency}"):
markets.append(result["symbol"])
return markets
def process_screener_data(app, markets, quote_currency, exchange_name):
"""
Hit TradingView up for the goods so we don't waste unnecessary time/compute resources (brandon's top picks)
"""
ta_screener_list = [f"{re.sub('PRO', '', app.exchange.name, re.IGNORECASE)}:{re.sub('-', '', market)}" for market in markets]
screener_staging = [p for p in chunker(ta_screener_list, 100)]
screener_analysis = []
additional_indicators = ["ATR", "KltChnl.upper", "KltChnl.lower"]
# TradingView.indicators.append("Volatility.D")
for pair_list in screener_staging:
screener_analysis.extend(
[
a
for a in get_multiple_analysis( # noqa: F405
screener="crypto", interval=app.granularity.short, symbols=pair_list, additional_indicators=additional_indicators
).values()
]
)
# Take what we need and do magic, ditch the rest.
formatted_ta = []
for ta in screener_analysis:
try:
if app.debug:
print(f"Checking {ta.symbol} on {exchange_name}\n")
recommend = Decimal(ta.indicators.get("Recommend.All"))
volatility = Decimal(
volatility_calculator(
ta.indicators["BB.upper"],
ta.indicators["BB.lower"],
ta.indicators["KltChnl.upper"],
ta.indicators["KltChnl.lower"],
ta.indicators["high"],
ta.indicators["low"],
)
)
# volatility = Decimal(ta.indicators['Volatility.D']) * 100
adx = abs(Decimal(ta.indicators["ADX"]))
adx_posi_di = Decimal(ta.indicators["ADX+DI"])
adx_neg_di = Decimal(ta.indicators["ADX-DI"])
high = Decimal(ta.indicators["high"]).quantize(Decimal("1e-{}".format(8))) # noqa: F841
low = Decimal(ta.indicators["low"]).quantize(Decimal("1e-{}".format(8))) # noqa: F841
close = Decimal(ta.indicators["close"]).quantize(Decimal("1e-{}".format(8)))
# ATR normalised
atr = (Decimal(ta.indicators["ATR"]) / close * 100).quantize(Decimal("1e-{}".format(2))) if "ATR" in ta.indicators else 0
volume = Decimal(ta.indicators["volume"])
macd = Decimal(ta.indicators["MACD.macd"])
macd_signal = Decimal(ta.indicators["MACD.signal"])
bollinger_upper = Decimal(ta.indicators["BB.upper"])
bollinger_lower = Decimal(ta.indicators["BB.lower"])
kelt_upper = Decimal(ta.indicators["KltChnl.upper"]) # noqa: F841
kelt_lower = Decimal(ta.indicators["KltChnl.lower"]) # noqa: F841
rsi = Decimal(ta.indicators.get("RSI", 0))
stoch_d = Decimal(ta.indicators.get("Stoch.D", 0))
stoch_k = Decimal(ta.indicators.get("Stoch.K", 0))
williams_r = Decimal(ta.indicators.get("W.R", 0))
score = 0
analysis_summary = ta.summary # noqa: F841
rating = ta.summary["RECOMMENDATION"]
# print(close)
if rating == "SELL":
score -= 2.5
elif rating == "STRONG_SELL":
score -= 5
elif rating == "NEUTRAL":
score += 0
elif rating == "BUY":
score += 2.5
elif rating == "STRONG_BUY":
score += 5
if (adx >= app.adx_threshold) and (adx_posi_di > adx_neg_di) and (adx_posi_di > adx):
if app.debug:
print(f"ADX({adx}) >= {app.adx_threshold}")
score += 1
if volume >= app.volume_threshold:
if app.debug:
print(f"Volume({volume}) >= {app.volume_threshold}")
score += 1
if abs(macd) > abs(macd_signal):
if app.debug:
print(f"MACD({macd}) above signal({macd_signal})")
score += 1
if volatility >= app.volatility_threshold:
if app.debug:
print(f"Volatility({volatility} is above {app.volatility_threshold}")
score += 1
if volatility < app.minimum_volatility:
if app.debug:
print(f"{ta.symbol} ({volatility}) is below min volatility of {app.minimum_volatility}")
score -= 100
if volume < app.minimum_volume:
if app.debug:
print(f"{ta.symbol} ({volume}) is below min volume of {app.volume}")
score -= 100
if close < app.minimum_quote_price:
if app.debug:
print(f"{ta.symbol} ({close}) is below min quote self.price of {app.minimum_quote_price}")
score -= 100
if 30 >= rsi > 20:
score += 1
if 20 < stoch_d <= 30:
score += 1
if stoch_k > stoch_d:
score += 1
if williams_r <= -30:
score += 1
# print('symbol\tscore\tvolume\tvvolatilith\tadx\tadx_posi_di\tadx_neg_di\tmacd\tmacd_signal\tbollinger_upper\tbollinger_lower\trecommend')
# print(ta.symbol, score, volume, volatility, adx, adx_posi_di, adx_neg_di, macd, macd_signal, bollinger_upper, bollinger_lower, recommend, "\n")
# print(f"Symbol: {ta.symbol} Score: {score}/{self.selection_score} Rating: {rating}")
if (score >= app.selection_score) and (rating in app.tv_screener_ratings):
relavent_ta = {}
if app.exchange == CryptoExchange.COINBASE or app.exchange == CryptoExchange.COINBASEPRO or app.exchange == CryptoExchange.KUCOIN:
relavent_ta["market"] = re.sub(rf"(.*){quote_currency}", rf"\1-{quote_currency}", ta.symbol)
# relavent_ta['market'] = re.sub(quote_currency,f"-{quote_currency}", ta.symbol)
else:
relavent_ta["market"] = ta.symbol
# relavent_ta['market'] = ta.symbol
relavent_ta["recommend"] = recommend
relavent_ta["volume"] = volume
relavent_ta["volatility"] = volatility
relavent_ta["adx"] = adx
relavent_ta["adx+di"] = adx_posi_di
relavent_ta["adx-di"] = adx_neg_di
relavent_ta["macd"] = macd
relavent_ta["macd.signal"] = macd_signal
relavent_ta["bollinger_upper"] = bollinger_upper
relavent_ta["bollinger_lower"] = bollinger_lower
relavent_ta["rsi"] = rsi
relavent_ta["stoch_d"] = stoch_d
relavent_ta["stoch_k"] = stoch_k
relavent_ta["williamsr"] = williams_r
relavent_ta["rating"] = rating
relavent_ta["score"] = score
## Hack a percentage from the recommendation which would take into account all the indicators rather than just ATR
if atr > 0:
relavent_ta["atr72_pcnt"] = atr
# else:
# if recommend > 0:
# relavent_ta['atr72_pcnt'] = recommend * 100
else:
relavent_ta["atr72_pcnt"] = 0
try:
relavent_ta["buy_next"] = "SEND IT!" if re.search("BUY", rating) else False
except AttributeError:
relavent_ta["buy_next"] = False
formatted_ta.append(relavent_ta)
except Exception:
pass
if formatted_ta:
# Stick it in a DF for the bots
df_markets = pd.DataFrame(formatted_ta)
df_markets = df_markets[
[
"market",
"score",
"recommend",
"volume",
"volatility",
"adx",
"adx+di",
"adx-di",
"macd",
"macd.signal",
"bollinger_upper",
"bollinger_lower",
"rsi",
"stoch_d",
"stoch_k",
"williamsr",
"rating",
"buy_next",
"atr72_pcnt",
]
]
df_markets.columns = [
"market",
"score",
"recommend",
"volume",
"volatility",
"adx",
"adx+di",
"adx-di",
"macd",
"macd.signal",
"bollinger_upper",
"bollinger_lower",
"rsi",
"stoch_d",
"stoch_k",
"williamsr",
"rating",
"buy_next",
"atr72_pcnt",
]
df_markets["score"] = df_markets["score"].astype(float).round(0).astype(int)
df_markets["recommend"] = df_markets["recommend"].astype(float)
df_markets["volume"] = df_markets["volume"].astype(float).round(0).astype(int)
df_markets["volatility"] = df_markets["volatility"].astype(float)
df_markets["adx"] = df_markets["adx"].astype(float)
df_markets["adx+di"] = df_markets["adx+di"].astype(float)
df_markets["adx-di"] = df_markets["adx-di"].astype(float)
df_markets["macd"] = df_markets["macd"].astype(float)
df_markets["macd.signal"] = df_markets["macd.signal"].astype(float)
df_markets["bollinger_upper"] = df_markets["bollinger_upper"].astype(float)
df_markets["bollinger_lower"] = df_markets["bollinger_lower"].astype(float)
df_markets["rsi"] = df_markets["rsi"].astype(float)
df_markets["stoch_d"] = df_markets["stoch_d"].astype(float)
df_markets["stoch_k"] = df_markets["stoch_k"].astype(float)
df_markets["williamsr"] = df_markets["williamsr"].astype(float)
df_markets["atr72_pcnt"] = df_markets["atr72_pcnt"].astype(float)
df_markets.sort_values(by=["market"], ascending=True, inplace=True)
df_markets.set_index("market", inplace=True)
print(df_markets.sort_values(by=["buy_next", "atr72_pcnt"], ascending=[False, False], inplace=False))
TGBot(app, scanner=True).save_scanner_output(app.exchange.value, quote_currency, df_markets)
else:
blank_data = {}
blank_data["buy_next"] = False
blank_data["atr72_pcnt"] = 0
blank_data["volume"] = 0
formatted_ta.append(blank_data)
df_markets = pd.DataFrame(formatted_ta)
TGBot(app, scanner=True).save_scanner_output(app.exchange.value, quote_currency, df_markets)
print("No pairs found!")
return True
if __name__ == "__main__":
tvlib_ver = version("tradingview-ta")
if tvlib_ver >= "3.2.10":
print(f"Library is correct version - were good to go! (v {tvlib_ver})")
else:
print(f"Gotta update your tradingview-ta library please! (v {tvlib_ver})")
sys.exit()
start_time = time.time()
print("Processing, please wait...")
bootstrap_exchanges = load_configs()
for app in bootstrap_exchanges:
print(f"\n\n{app.exchange.name}")
for quote_currency in app.scanner_quote_currencies:
markets = get_markets(app, quote_currency)
try:
process_screener_data(app, markets, quote_currency, app.exchange.name)
except Exception as e:
print(e)
print("Scan run finished!")
print(f"Timestamp: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}")
print(f"Total elapsed time: {time.time() - start_time} sec")