FEAT: support multiple tickers

Author: gialdetti

CHANGELOG.md

@@ -6,7 +6,9 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
 ## [Unreleased]
+### Added
 * MAINT: update setup.py to support multiple modes
+* API: support multiple tickers
 ### Fixed
 * Align with the pandas new APIs.
 

capon/__init__.py

@@ -1,24 +1,23 @@
-from .backends.yahoo import stock
+import logging
+from .wrappers import stock, stocks
 from .backends.nasdaq import metadata
 
+
+logger = logging.getLogger(__name__)
+
+
 try:
     from .visualization import template
 except ModuleNotFoundError as e:
-    import logging
-
-    logger = logging.getLogger(__name__)
     logger.warning(e)
 
 try:
     # from .visualization import plot_history as plot
     from .visualization.altairplotter import plot_history as plot
 except ModuleNotFoundError as e:
-    import logging
-
-    logger = logging.getLogger(__name__)
     logger.warning(e)
 
     def plot(*args, **kwargs):
         raise ImportError(
-            "Missing optional dependency 'altair'.  Use pip or conda to install it."
+            "Missing optional dependency 'altair'. Use pip or conda to install it."
         )

capon/backends/utils.py

@@ -2,36 +2,69 @@
 import re
 import requests
 
+import pandas as pd
+
+
+#
+# Requests
+#
+
 
 def get_json(url, headers={}):
     response = requests.get(url, headers=headers)
-    logging.info(f'{response.url}.. {response.status_code}')
+    logging.info(f"{response.url}.. {response.status_code}")
 
     jo = {}
-    if response.status_code==200:
+    if response.status_code == 200:
         jo = response.json()
 
     return jo
 
 
 def camel_case_keys(df):
-    last_dim = len(df.shape)-1
-    df = df.rename({k: camel_to_snake_case(k) for k in df.keys()}, axis=last_dim) \
-        .rename({
-            'forward_pe1_yr': 'forward_pe_1_yr',
-            'is_nasdaq100': 'is_nasdaq_100',
-        }, axis=last_dim)
+    last_dim = len(df.shape) - 1
+    df = df.rename(
+        {k: camel_to_snake_case(k) for k in df.keys()}, axis=last_dim
+    ).rename(
+        {
+            "forward_pe1_yr": "forward_pe_1_yr",
+            "is_nasdaq100": "is_nasdaq_100",
+        },
+        axis=last_dim,
+    )
     return df
 
 
 def camel_to_snake_case(name):
-    name = re.sub(' ', '', name)
-    name = re.sub('(.)([A-Z][a-z]+)', r'\1_\2', name)
-    name = re.sub('([a-z0-9])([A-Z])', r'\1_\2', name).lower()
+    name = re.sub(" ", "", name)
+    name = re.sub("(.)([A-Z][a-z]+)", r"\1_\2", name)
+    name = re.sub("([a-z0-9])([A-Z])", r"\1_\2", name).lower()
     return name
 
 
+#
+# Pandas
+#
+
+
+def reindex(df: pd.DataFrame, moved: list, position: str = "start", axis: int = 1):
+    assert axis == 1
+
+    rest = [c for c in df if c not in moved]
+    if position == "start":
+        new_labels = moved + rest
+    elif position == "end":
+        new_labels = rest + moved
+    else:
+        # can support int position too
+        raise ValueError()
+
+    df_reindexed = df.reindex(new_labels, axis=axis)
+
+    return df_reindexed
+
+
 if False:
-    camel_to_snake_case('getHTTPResponseCode') == 'get_http_response_code'
-    camel_to_snake_case('ForwardPE1Yr') == 'forward_pe1_yr'
-    camel_to_snake_case('Security Name') == 'security_name'
+    camel_to_snake_case("getHTTPResponseCode") == "get_http_response_code"
+    camel_to_snake_case("ForwardPE1Yr") == "forward_pe1_yr"
+    camel_to_snake_case("Security Name") == "security_name"

capon/backends/yahoo.py

@@ -8,6 +8,7 @@
 from datetime import datetime
 
 import pandas as pd
+from capon.backends import utils
 
 
 logger = logging.getLogger(__name__)
@@ -72,7 +73,7 @@ def get_stock(symbol, range="1d", interval=None, start_date=None, end_date=None)
     return jo
 
 
-def stock(symbol, range="1d", interval=None, start_date=None, end_date=None):
+def stock(symbol, range="1d", interval=None, start=None, end=None):
     """Get live & historical stock prices.
 
     Parameters
@@ -92,7 +93,7 @@ def stock(symbol, range="1d", interval=None, start_date=None, end_date=None):
         low, and close price for a given timepoint. If relevant, it will also include the adjusted closing price.
     """
     jo = get_stock(
-        symbol, range=range, interval=interval, start_date=start_date, end_date=end_date
+        symbol, range=range, interval=interval, start_date=start, end_date=end
     )
 
     result = jo["chart"]["result"][0]
@@ -114,7 +115,11 @@ def stock(symbol, range="1d", interval=None, start_date=None, end_date=None):
     )
 
     # len(ts), len(quote), len(adjclose)
-    stock = pd.concat([ts, quote, adjclose], axis=1)
+    stock = pd.concat([ts, quote, adjclose], axis=1).pipe(
+        utils.reindex,
+        ["timestamp", "symbol", "currency", "open", "low", "high", "close", "volume"],
+    )
+
     return stock
 
 

capon/portfolio.py

@@ -220,7 +220,7 @@ def fetch_history(self, symbols=None, start_time=None, end_time=None, tz=None):
         history = pd.concat(
             [
                 capon.stock(
-                    symbol, start_date=start_time, end_date=end_time, interval="1d"
+                    symbol, start=start_time, end=end_time, interval="1d"
                 ).dropna()
                 for symbol in tqdm(symbols)
             ],

capon/tests/test_core.py

@@ -1,19 +1,110 @@
+import pandas.testing as pdt
+
 import capon
 
 
 def test_adjclose():
-    amd_daily = capon.stock('AMD', range='1mo', interval='1d')
-    assert 'adjclose' in amd_daily
+    amd_daily = capon.stock("AMD", range="1mo", interval="1d")
+    assert "adjclose" in amd_daily
 
-    amd_minutely = capon.stock('AMD', range='1d', interval='2m')
-    assert 'close' in amd_minutely
+    amd_minutely = capon.stock("AMD", range="1d", interval="2m")
+    assert "close" in amd_minutely
 
 
 def test_index():
-    sp500 = capon.stock('^GSPC')
+    sp500 = capon.stock("^GSPC")
     assert len(sp500) > 0
 
 
 def test_metadata():
-    metadata = capon.metadata('AMD')
-    assert metadata['symbol'] == 'AMD'
+    metadata = capon.metadata("AMD")
+    assert metadata["symbol"] == "AMD"
+
+
+if False:
+    import logging
+
+    from IPython.display import display
+
+    logging.basicConfig()
+    logging.getLogger("capon").setLevel(logging.DEBUG)
+
+
+base_tickers = ["MSFT", "AAPL", "NVDA"]
+
+
+def test_stocks():
+    capon.stocks(["MSFT", "AAPL", "NVDA"])
+
+
+def test_stocks_njobs():
+    all_quotes = []
+    for n_jobs in [None, -1]:
+        quotes = capon.stocks(
+            base_tickers,
+            start="2020-01-01",
+            end="2021-01-01",
+            interval="1d",
+            n_jobs=n_jobs,
+        )
+        all_quotes.append(quotes)
+
+    pdt.assert_frame_equal(all_quotes[0], all_quotes[1])
+
+
+def test_stocks_missing():
+    missing_ticker = "BZZ"
+
+    quotes = capon.stocks(
+        base_tickers + [missing_ticker],
+        start="2020-01-01",
+        end="2021-01-01",
+        interval="1d",
+    )
+    # display(quotes)
+    assert (quotes["symbol"].unique() == base_tickers).all()
+
+
+def test_stocks_multiple_tz():
+    international_ticker = "^TA125.TA"
+
+    for timestamp_normalizer in ["auto", None, "date"]:
+        quotes = capon.stocks(
+            base_tickers + [international_ticker],
+            start="2020-01-01",
+            end="2021-01-01",
+            interval="1d",
+            timestamp_normalizer=timestamp_normalizer,
+        )
+        # display(quotes)
+
+        assert (
+            quotes["symbol"].unique() == base_tickers + [international_ticker]
+        ).all()
+
+
+def test_stocks_single_ticker():
+    quotes = capon.stocks(
+        base_tickers[0],
+        start="2020-01-01",
+        end="2021-01-01",
+        interval="1d",
+    )
+
+    assert (quotes["symbol"].unique() == base_tickers[0]).all()
+
+
+# def test_stocks_timestamp_auto():
+#     quotes_1h = capon.stocks(
+#         base_tickers,
+#         range="1mo",
+#         interval="1h",
+#         timestamp_normalizer="auto",
+#     )
+
+#     quotes_1d = capon.stocks(
+#         base_tickers,
+#         range="1mo",
+#         interval="1d",
+#         timestamp_normalizer="auto",
+#     )

capon/tests/visualization/test_template.py

@@ -1,6 +1,10 @@
-import plotly.io as pio
 import capon
 
+try:
+    import plotly.io as pio
 
-def test_template_registration():
-    assert 'capon' in pio.templates
+    def test_template_registration():
+        assert "capon" in pio.templates
+
+except ModuleNotFoundError as e:
+    print(e)

capon/visualization/template.py

@@ -3,29 +3,21 @@
 import plotly.express as px
 
 
-pio.templates['capon'] = go.layout.Template(
+pio.templates["capon"] = go.layout.Template(
     layout=go.Layout(
-        font=dict(color='#887e7d'),
-        paper_bgcolor='#fff1e5',
-        plot_bgcolor='#fff1e5',
-
+        font=dict(color="#887e7d"),
+        paper_bgcolor="#fff1e5",
+        plot_bgcolor="#fff1e5",
         title=dict(
-            font=dict(size=22, color='black', family='Palatino'),
-            xanchor='left',
-            x=0.05),
-
+            font=dict(size=22, color="black", family="Palatino"), xanchor="left", x=0.05
+        ),
         colorway=px.colors.qualitative.T10,
-
         colorscale={
-            'sequential': 'Geyser_r',
-            'sequentialminus': 'Geyser_r',
-        }
+            "sequential": "Geyser_r",
+            "sequentialminus": "Geyser_r",
+        },
     ),
-
-    data_scatter=[
-        go.Scatter(line=dict(width=2.5))
-    ],
-
+    data_scatter=[go.Scatter(line=dict(width=2.5))],
 )
 
 # https://plotly.com/python/builtin-colorscales/
@@ -34,4 +26,4 @@
 
 # from plotly.offline import plot
 # fig = px.line(x=[1,2,3], y=[1,3,2], title='Test', template='capon')
-# plot(fig)
+# plot(fig)