identify_trends function for investpy developed

.gitignore

@@ -102,3 +102,8 @@ venv.bak/
 
 # mypy
 .mypy_cache/
+
+# PyCharm IDEA files
+.idea
+.idea/
+.idea/*

trendet/__init__.py

@@ -0,0 +1,135 @@
+#!/usr/bin/env python
+
+# Copyright 2018-2019 Alvaro Bartolome
+# See LICENSE for details.
+
+__author__ = 'Alvaro Bartolome @ alvarob96 on GitHub'
+__version__ = '0.1'
+
+from investpy import get_historical_data
+
+from statistics import mean
+import datetime
+import string
+
+
+def identify_trends(equity, from_date, to_date, window_size=5, trend_limit=3, labels=None):
+    """
+    This function retrieves historical data from the introduced `equity` from Investing
+    via Web Scraping on the introduced date range. The resulting data can it either be
+    stored in a :obj:`pandas.DataFrame` or in a :obj:`json` object with `ascending` or `descending` order.
+
+    Args:
+        equity (:obj:`str`): name of the equity to retrieve historical data from.
+        from_date (:obj:`str`): date as `str` formatted as `dd/mm/yyyy`, from where data is going to be retrieved.
+        to_date (:obj:`str`): date as `str` formatted as `dd/mm/yyyy`, until where data is going to be retrieved.
+        window_size (:obj:`window`, optional): number of days from where market behaviour is considered a trend.
+        trend_limit (:obj:`int`, optional): maximum number of trends to identify
+        labels (:obj:`list`, optional): name of the labels for every identified trend.
+
+    Returns:
+        :obj:`pandas.DataFrame`:
+            The function returns a :obj:`pandas.DataFrame` which contains the retrieved historical data from Investing
+            using investpy, with a new column which identifies every trend found on the market between two dates
+            identifying when did the trend started and when did it end. So the additional column contains labeled date
+            ranges.
+    Raises:
+        ValueError: argument error.
+    """
+
+    if not isinstance(equity, str):
+        raise ValueError("equity argument needs to be a str.")
+
+    if not equity:
+        raise ValueError("equity parameter is mandatory and must be a valid equity name.")
+
+    try:
+        datetime.datetime.strptime(from_date, '%d/%m/%Y')
+    except ValueError:
+        raise ValueError("incorrect from_date date format, it should be 'dd/mm/yyyy'.")
+
+    try:
+        datetime.datetime.strptime(to_date, '%d/%m/%Y')
+    except ValueError:
+        raise ValueError("incorrect to_date format, it should be 'dd/mm/yyyy'.")
+
+    start_date = datetime.datetime.strptime(from_date, '%d/%m/%Y')
+    end_date = datetime.datetime.strptime(to_date, '%d/%m/%Y')
+
+    if start_date >= end_date:
+        raise ValueError("to_date should be greater than from_date, both formatted as 'dd/mm/yyyy'.")
+
+    if not isinstance(window_size, int):
+        raise ValueError('window_size must be an `int`')
+
+    if isinstance(window_size, int) and window_size < 3:
+        raise ValueError('window_size must be an `int` equal or higher than 3!')
+
+    if not isinstance(trend_limit, int):
+        raise ValueError('trend_limit must be an `int`')
+
+    if isinstance(trend_limit, int) and trend_limit < 1:
+        raise ValueError('trend_limit must be an `int` equal or higher than 1!')
+
+    if labels is not None and isinstance(labels, list):
+        if len(labels) != len(trend_limit):
+            raise ValueError('if labels is not None and a `list`, it must have the same length as the trend_limit!')
+
+    if labels is not None and not isinstance(labels, list):
+        raise ValueError('labels is neither None or a `list`!')
+
+    if labels is None:
+        labels = [letter for letter in string.ascii_uppercase[:trend_limit]]
+
+    try:
+        df = get_historical_data(equity=str(equity),
+                                 from_date=from_date,
+                                 to_date=to_date,
+                                 as_json=False,
+                                 order='ascending',
+                                 debug=False)
+    except:
+        raise RuntimeError('investpy function call failed!')
+
+    limit = None
+    values = list()
+
+    trends = list()
+
+    for index, value in enumerate(df['Close'], 0):
+        if limit and limit > value:
+            values.append(value)
+            limit = mean(values)
+        elif limit and limit < value:
+            if len(values) > window_size:
+                min_value = min(values)
+
+                for counter, item in enumerate(values, 0):
+                    if item == min_value:
+                        break
+
+                to_trend = from_trend + counter
+
+                obj = {
+                    'from': df.index.tolist()[from_trend],
+                    'to': df.index.tolist()[to_trend],
+                }
+
+                trends.append(obj)
+
+                if len(trends) >= trend_limit:
+                    break
+
+            limit = None
+            values = list()
+        else:
+            from_trend = index
+
+            values.append(value)
+            limit = mean(values)
+
+    for trend, label in zip(trends, labels):
+        for index, row in df[trend['from']:trend['to']].iterrows():
+            df.loc[index, 'Trend'] = label
+
+    return df