Merge pull request #2637 from ranaroussi/dev

sync dev -> main

Author: ValueRaider

doc/source/advanced/config.rst

@@ -2,14 +2,55 @@
 Config
 ******
 
-`yfinance` has a new global config for sharing common values.
+`yfinance` has a new global config for sharing common values:
 
-Proxy
+.. code-block:: python
+
+  >>> import yfinance as yf
+  >>> yf.config
+  {
+    "network": {
+      "proxy": null,
+      "retries": 0
+    },
+    "debug": {
+      "hide_exceptions": true,
+      "logging": false
+    }
+  }
+  >>> yf.config.network
+  {
+    "proxy": null,
+    "retries": 0
+  }
+
+
+Network
+-------
+
+* **proxy** - Set proxy for all yfinance data fetches.
+
+  .. code-block:: python
+
+     yf.config.network.proxy = "PROXY_SERVER"
+
+* **retries** - Configure automatic retry for transient network errors. The retry mechanism uses exponential backoff (1s, 2s, 4s...).
+
+  .. code-block:: python
+
+     yf.config.network.retries = 2
+
+Debug
 -----
 
-Set proxy once in config, affects all yfinance data fetches.
+* **hide_exceptions** - Set to `False` to stop yfinance hiding exceptions.
 
-.. code-block:: python
+  .. code-block:: python
+
+     yf.config.debug.hide_exceptions = False
+
+* **logging** - Set to `True` to enable verbose debug logging.
+
+  .. code-block:: python
 
-   import yfinance as yf
-   yf.set_config(proxy="PROXY_SERVER")
+     yf.config.debug.logging = True

doc/source/reference/examples/calendars.py

@@ -0,0 +1,38 @@
+import yfinance as yf
+from datetime import datetime, timedelta
+
+# Default init (today + 7 days)
+calendar = yf.Calendars()
+
+# Today's events: calendar of 1 day
+tomorrow = datetime.now() + timedelta(days=1)
+calendar = yf.Calendars(end=tomorrow)
+
+# Default calendar queries - accessing the properties will fetch the data from YF
+calendar.earnings_calendar
+calendar.ipo_info_calendar
+calendar.splits_calendar
+calendar.economic_events_calendar
+
+# Manual queries
+calendar.get_earnings_calendar()
+calendar.get_ipo_info_calendar()
+calendar.get_splits_calendar()
+calendar.get_economic_events_calendar()
+
+# Earnings calendar custom filters
+calendar.get_earnings_calendar(
+    market_cap=100_000_000,  # filter out small-cap 
+    filter_most_active=True,  # show only actively traded. Uses: `screen(query="MOST_ACTIVES")`
+)
+
+# Example of real use case:
+# Get inminent unreported earnings events
+today = datetime.now()
+is_friday = today.weekday() == 4
+day_after_tomorrow = today + timedelta(days=4 if is_friday else 2)
+
+calendar = yf.Calendars(today, day_after_tomorrow)
+df = calendar.get_earnings_calendar(limit=100)
+
+unreported_df = df[df["Reported EPS"].isnull()]

doc/source/reference/index.rst

@@ -16,6 +16,7 @@ The following are the publicly available classes, and functions exposed by the `
 - :attr:`Ticker <yfinance.Ticker>`: Class for accessing single ticker data.
 - :attr:`Tickers <yfinance.Tickers>`: Class for handling multiple tickers.
 - :attr:`Market <yfinance.Market>`: Class for accessing market summary.
+- :attr:`Calendars <yfinance.Calendars>`: Class for accessing calendar events data.
 - :attr:`download <yfinance.download>`: Function to download market data for multiple tickers.
 - :attr:`Search <yfinance.Search>`: Class for accessing search results.
 - :attr:`Lookup <yfinance.Lookup>`: Class for looking up tickers.
@@ -37,6 +38,7 @@ The following are the publicly available classes, and functions exposed by the `
    yfinance.ticker_tickers
    yfinance.stock
    yfinance.market
+   yfinance.calendars
    yfinance.financials
    yfinance.analysis
    yfinance.search

doc/source/reference/yfinance.calendars.rst

@@ -0,0 +1,21 @@
+=====================
+Calendars
+=====================
+
+.. currentmodule:: yfinance
+
+
+Class
+------------
+The `Calendars` class allows you to get information about upcoming events, for example, earning events.
+
+.. autosummary::
+   :toctree: api/
+
+   Calendars
+
+Sample Code
+------------------
+
+.. literalinclude:: examples/calendars.py
+   :language: python

meta.yaml

@@ -15,6 +15,7 @@ build:
   script: "{{ PYTHON }} -m pip install . --no-deps --ignore-installed -vv "
 
 requirements:
+  # curl_cffi 0.14 has major problems, see their Github
   host:
     - pandas >=1.3.0
     - numpy >=1.16.5
@@ -26,7 +27,7 @@ requirements:
     - frozendict >=2.3.4
     - beautifulsoup4 >=4.11.1
     - html5lib >=1.1
-    - curl_cffi >=0.7
+    - curl_cffi >=0.7,<0.14
     - peewee >=3.16.2
     - pip
     - python
@@ -42,7 +43,7 @@ requirements:
     - frozendict >=2.3.4
     - beautifulsoup4 >=4.11.1
     - html5lib >=1.1
-    - curl_cffi >=0.7
+    - curl_cffi >=0.7,<0.14
     - peewee >=3.16.2
     - python
 

requirements.txt

@@ -10,6 +10,7 @@ peewee>=3.16.2
 requests_cache>=1.0
 requests_ratelimiter>=0.3.1
 scipy>=1.6.3
-curl_cffi>=0.7
+# curl_cffi 0.14 has major problems, see their Github
+curl_cffi>=0.7,<0.14
 protobuf>=3.19.0
 websockets>=13.0

setup.py

@@ -59,11 +59,12 @@
     platforms=['any'],
     keywords='pandas, yahoo finance, pandas datareader',
     packages=find_packages(exclude=['contrib', 'docs', 'tests', 'examples']),
+    # curl_cffi 0.14 has major problems, see their Github
     install_requires=['pandas>=1.3.0', 'numpy>=1.16.5',
                       'requests>=2.31', 'multitasking>=0.0.7',
                       'platformdirs>=2.0.0', 'pytz>=2022.5',
                       'frozendict>=2.3.4', 'peewee>=3.16.2',
-                      'beautifulsoup4>=4.11.1', 'curl_cffi>=0.7',
+                      'beautifulsoup4>=4.11.1', 'curl_cffi>=0.7,<0.14',
                       'protobuf>=3.19.0', 'websockets>=13.0'],
     extras_require={
         'nospam': ['requests_cache>=1.0', 'requests_ratelimiter>=0.3.1'],

tests/test_calendars.py

@@ -0,0 +1,55 @@
+from datetime import datetime, timedelta, timezone
+import unittest
+
+import pandas as pd
+
+from tests.context import yfinance as yf, session_gbl
+
+
+class TestCalendars(unittest.TestCase):
+    def setUp(self):
+        self.calendars = yf.Calendars(session=session_gbl)
+
+    def test_get_earnings_calendar(self):
+        result = self.calendars.get_earnings_calendar(limit=1)
+        tickers = self.calendars.earnings_calendar.index.tolist()
+
+        self.assertIsInstance(result, pd.DataFrame)
+        self.assertEqual(len(result), 1)
+        self.assertIsInstance(tickers, list)
+        self.assertEqual(len(tickers), len(result))
+        self.assertEqual(tickers, result.index.tolist())
+        
+        first_ticker = result.index.tolist()[0]
+        result_first_ticker = self.calendars.earnings_calendar.loc[first_ticker].name
+        self.assertEqual(first_ticker, result_first_ticker)
+
+    def test_get_earnings_calendar_init_params(self):
+        result = self.calendars.get_earnings_calendar(limit=5)
+        self.assertGreaterEqual(result['Event Start Date'].iloc[0], pd.to_datetime(datetime.now(tz=timezone.utc)))
+
+        start = datetime.now(tz=timezone.utc) - timedelta(days=7)
+        result = yf.Calendars(start=start).get_earnings_calendar(limit=5)
+        self.assertGreaterEqual(result['Event Start Date'].iloc[0], pd.to_datetime(start))
+
+    def test_get_ipo_info_calendar(self):
+        result = self.calendars.get_ipo_info_calendar(limit=5)
+
+        self.assertIsInstance(result, pd.DataFrame)
+        self.assertEqual(len(result), 5)
+
+    def test_get_economic_events_calendar(self):
+        result = self.calendars.get_economic_events_calendar(limit=5)
+
+        self.assertIsInstance(result, pd.DataFrame)
+        self.assertEqual(len(result), 5)
+
+    def test_get_splits_calendar(self):
+        result = self.calendars.get_splits_calendar(limit=5)
+
+        self.assertIsInstance(result, pd.DataFrame)
+        self.assertEqual(len(result), 5)
+
+
+if __name__ == "__main__":
+    unittest.main()

tests/test_price_repair.py

@@ -61,7 +61,7 @@ def test_resampling(self):
                         vol_diff_pct0 = (dfr['Volume'].iloc[0] - df_truth['Volume'].iloc[0])/df_truth['Volume'].iloc[0]
                         vol_diff_pct1 = (dfr['Volume'].iloc[-1] - df_truth['Volume'].iloc[-1])/df_truth['Volume'].iloc[-1]
                         vol_diff_pct = _np.array([vol_diff_pct0, vol_diff_pct1])
-                        vol_match = vol_diff_pct > -0.23
+                        vol_match = vol_diff_pct > -0.32
                         vol_match_nmatch = _np.sum(vol_match)
                         vol_match_ndiff = len(vol_match) - vol_match_nmatch
                         if vol_match.all():
@@ -78,6 +78,8 @@ def test_resampling(self):
 
                     if debug:
                         print("- investigate:")
+                        print(f"  - interval = {interval}")
+                        print(f"  - period = {period}")
                         print("- df_truth:")
                         print(df_truth)#[['Open', 'Close', 'Volume']])
                         df_1d = dat.history(interval='1d', period=period)
@@ -361,27 +363,23 @@ def test_repair_zeroes_daily(self):
         hist = dat._lazy_load_price_history()
         tz_exchange = dat.fast_info["timezone"]
 
-        df_bad = _pd.DataFrame(data={"Open":      [0,      114.37, 114.20],
-                                     "High":      [0,      114.40, 114.40],
-                                     "Low":       [0,      114.36, 114.20],
-                                     "Close":     [114.39, 114.38, 114.45],
-                                     "Adj Close": [114.39, 114.38, 114.45],
-                                     "Volume":    [9,      15666,  1094]},
-                               index=_pd.to_datetime([_dt.datetime(2025, 3, 17),
-                                                      _dt.datetime(2025, 3, 14),
-                                                      _dt.datetime(2025, 3, 13)]))
-        df_bad = df_bad.sort_index()
-        df_bad.index.name = "Date"
-        df_bad.index = df_bad.index.tz_localize(tz_exchange)
+        correct_df = dat.history(period='1mo', auto_adjust=False)
+
+        dt_bad = correct_df.index[len(correct_df)//2]
+        df_bad = correct_df.copy()
+        for c in df_bad.columns:
+            df_bad.loc[dt_bad, c] = _np.nan
 
         repaired_df = hist._fix_zeroes(df_bad, "1d", tz_exchange, prepost=False)
 
-        correct_df = df_bad.copy()
-        correct_df.loc["2025-03-17", "Open"] = 114.62
-        correct_df.loc["2025-03-17", "High"] = 114.62
-        correct_df.loc["2025-03-17", "Low"] = 114.41
         for c in ["Open", "Low", "High", "Close"]:
-            self.assertTrue(_np.isclose(repaired_df[c], correct_df[c], rtol=1e-7).all())
+            try:
+                self.assertTrue(_np.isclose(repaired_df[c], correct_df[c], rtol=1e-7).all())
+            except Exception:
+                print(f"# column = {c}")
+                print("# correct:") ; print(correct_df[c])
+                print("# repaired:") ; print(repaired_df[c])
+                raise
 
         self.assertTrue("Repaired?" in repaired_df.columns)
         self.assertFalse(repaired_df["Repaired?"].isna().any())
@@ -421,7 +419,13 @@ def test_repair_zeroes_daily_adjClose(self):
 
                 df_slice_bad_repaired = hist._fix_zeroes(df_slice_bad, "1d", tz_exchange, prepost=False)
                 for c in ["Close", "Adj Close"]:
-                    self.assertTrue(_np.isclose(df_slice_bad_repaired[c], df_slice[c], rtol=rtol).all())
+                    try:
+                        self.assertTrue(_np.isclose(df_slice_bad_repaired[c], df_slice[c], rtol=rtol).all())
+                    except Exception:
+                        print(f"# column = {c}")
+                        print("# correct:") ; print(df_slice[c])
+                        print("# repaired:") ; print(df_slice_bad_repaired[c])
+                        raise
                 self.assertTrue("Repaired?" in df_slice_bad_repaired.columns)
                 self.assertFalse(df_slice_bad_repaired["Repaired?"].isna().any())
 
@@ -464,7 +468,7 @@ def test_repair_bad_stock_splits(self):
         # Stocks that split in 2022 but no problems in Yahoo data,
         # so repair should change nothing
         good_tkrs = ['AMZN', 'DXCM', 'FTNT', 'GOOG', 'GME', 'PANW', 'SHOP', 'TSLA']
-        good_tkrs += ['AEI', 'GHI', 'IRON', 'LXU', 'RSLS', 'TISI']
+        good_tkrs += ['AEI', 'GHI', 'IRON', 'LXU', 'TISI']
         good_tkrs += ['BOL.ST', 'TUI1.DE']
         intervals = ['1d', '1wk', '1mo', '3mo']
         for tkr in good_tkrs:
@@ -589,7 +593,6 @@ def test_repair_bad_div_adjusts(self):
 
         # Div 0.01x
         bad_tkrs += ['NVT.L']
-        bad_tkrs += ['TENT.L']
 
         # Missing div adjusts:
         bad_tkrs += ['1398.HK']

tests/test_prices.py

@@ -2,6 +2,7 @@
 from tests.context import session_gbl
 
 import unittest
+import socket
 
 import datetime as _dt
 import pytz as _tz
@@ -456,6 +457,21 @@ def test_aggregate_capital_gains(self):
 
         dat.history(start=start, end=end, interval=interval)
 
+    def test_transient_error_detection(self):
+        """Test that _is_transient_error correctly identifies transient vs permanent errors"""
+        from yfinance.data import _is_transient_error
+        from yfinance.exceptions import YFPricesMissingError
+
+        # Transient errors (should retry)
+        self.assertTrue(_is_transient_error(socket.error("Network error")))
+        self.assertTrue(_is_transient_error(TimeoutError("Timeout")))
+        self.assertTrue(_is_transient_error(OSError("OS error")))
+
+        # Permanent errors (should NOT retry)
+        self.assertFalse(_is_transient_error(ValueError("Invalid")))
+        self.assertFalse(_is_transient_error(YFPricesMissingError('INVALID', '')))
+        self.assertFalse(_is_transient_error(KeyError("key")))
+
 
 if __name__ == '__main__':
     unittest.main()