The Positive Similarity of Company Filings and Stock Returns

1from AlgoLib import *
2# endregion
3
4class ThePositiveSimilarityOfCompanyFilingsAndStockReturns(XXX):
5
6def Initialize(self):
7self.SetStartDate(2009, 1, 1) # first metric data come in 2009
8self.SetCash(100000)
9        
10self.leverage:int = 5
11self.quantile:int = 10
12
13self.metric_symbols:dict[Symbol, Symbol] = {}
14self.positive_similarities:dict[Symbol, float] = {}
15
16self.market:Symbol = self.AddEquity('SPY', Resolution.Daily).Symbol
17
18self.coarse_count:int = 1000
19self.pick_largest:bool = True
20
21self.selection_flag:bool = False
22self.UniverseSettings.Resolution = Resolution.Daily
23self.AddUniverse(self.CoarseSelectionFunction, self.FineSelectionFunction)
24self.Schedule.On(self.DateRules.MonthStart(self.market), self.TimeRules.BeforeMarketClose(self.market, 0), self.Selection)
25
26def OnSecuritiesChanged(self, changes):
27for security in changes.AddedSecurities:
28    security.SetFeeModel(CustomFeeModel())
29    security.SetLeverage(self.leverage)
30
31def CoarseSelectionFunction(self, coarse):
32if not self.selection_flag:
33    return Universe.Unchanged
34
35if self.coarse_count <= 1000 and not self.pick_largest:
36    selected:list = sorted([x for x in coarse if x.HasFundamentalData and x.Market == 'usa'],
37            key=lambda x: x.DollarVolume, reverse=True)[:self.coarse_count]
38else:
39    selected:list = [x for x in coarse if x.HasFundamentalData and x.Market == 'usa']
40
41selected_symbols:list[Symbol] = []
42
43for stock in selected:
44    symbol:Symbol = stock.Symbol
45
46    if symbol not in self.metric_symbols:
47        metric_symbol:Symbol = self.AddData(BrainCompanyFilingLanguageMetrics10K, symbol).Symbol
48        self.metric_symbols[symbol] = metric_symbol
49
50    selected_symbols.append(symbol)
51
52return selected_symbols
53
54def FineSelectionFunction(self, fine):
55if self.coarse_count <= 1000:
56    return list(map(lambda stock: stock.Symbol, fine))
57
58fine:list = list(filter(lambda stock: stock.MarketCap != 0, fine))
59
60if len(fine) > self.coarse_count or self.pick_largest:
61    sorted_by_cap:list = sorted(fine, key=lambda stock: stock.MarketCap)
62    fine = sorted_by_cap[-self.coarse_count:]
63
64return list(map(lambda stock: stock.Symbol, fine)) 
65
66def OnData(self, data):
67if self.selection_flag:
68    self.selection_flag = False
69
70    filtered_positive_similarity:dict[Symbol, float] = { symbol: pos_sim for symbol, pos_sim in self.positive_similarities.items() \
71        if symbol in data and data[symbol] }
72
73    self.positive_similarities.clear()
74
75    if len(filtered_positive_similarity) < self.quantile:
76        self.Liquidate()
77    else:
78        quantile:int = int(len(filtered_positive_similarity) / self.quantile)
79        sorted_by_pos_sim:list[Symbol] = [x[0] for x in sorted(filtered_positive_similarity.items(), key=lambda item: item[1])]
80
81        long_leg:list[Symbol] = sorted_by_pos_sim[:quantile]
82        short_leg:list[Symbol] = sorted_by_pos_sim[-quantile:]
83
84        invested:list[Symbol] = [x.Key for x in self.Portfolio if x.Value.Invested]
85        for symbol in invested:
86            if symbol not in long_leg + short_leg:
87                self.Liquidate(symbol)
88
89        for symbol in long_leg:
90            self.SetHoldings(symbol, 1 / quantile)
91
92        for symbol in short_leg:
93            self.SetHoldings(symbol, -1 / quantile)
94
95for stock_symbol, metric_symbol in self.metric_symbols.items():
96    if metric_symbol in data and data[metric_symbol]:
97        positive_similarity:float = data[metric_symbol].ReportSentiment.Similarity.Positive
98
99        if positive_similarity:
100            self.positive_similarities[stock_symbol] = positive_similarity
101
102def Selection(self):
103self.selection_flag = True
104
105# Custom fee model
106class CustomFeeModel(FeeModel):
107def GetOrderFee(self, parameters):
108fee = parameters.Security.Price * parameters.Order.AbsoluteQuantity * 0.00005
109return OrderFee(CashAmount(fee, "USD"))