从多个数据框中将多个图转换为pdf

Question

I would like to periodically create pdfs with stocks of interest to me and related graphs for each stock.

For the moment, I have one dataframe for each technical indicator (price, moving average, exponential moving average, etc.) and would like to plot for each time series their relevant information from each dataframe.

Until now, I have used bokeh and I am able to plot what I want but for each ticker individually...

The code I have is this:

# Import dataframes from import_data.py
prices = pd.read_pickle('prices.pkl')
date = prices.index
# price 20, 50, 80 MA
ma20 = pd.rolling_mean(prices, window=20)
ma50 = pd.rolling_mean(prices, window=50)
ma80 = pd.rolling_mean(prices, window=80)
# price with 20EMA+Bollinger Bands - 20MA
ema20 = pd.ewma(prices, span=20)
bbandsup = pd.ewma(prices, span=20) + 2 * pd.ewmstd(prices, span=20)
bbandslo = pd.ewma(prices, span=20) - 2 * pd.ewmstd(prices, span=20)


plot = figure(title="Price Chart and Technical Indicators", x_axis_label='Date', y_axis_label='Price')

plot.line(date, prices['SPY'], legend='Price', line_width=2)
plot.line(date, ema20['SPY'], legend='EMA-20', line_width=2, line_color="red")
plot.line(date, ma20['SPY'], legend='EA-20', line_width=2, line_color="green")
plot.line(date, bbandsup['SPY'], legend='Bollinger', line_width=2, line_color="black", line_dash="4 4")
plot.line(date, bbandslo['SPY'], legend='Bollinger', line_width=2, line_color="black", line_dash="4 4")

show(plot)

prices dataframe has a structure of date as index and the column headers are the ticker strings. For example, the first column header is SPY, the second is AAPL, etc...

I would like to be able to print all of them into pdf before attacking the multiple plot beast.

Thanks in advance!

Answer 1

Multiple plots are not more complicated than what you already have - you just have to call figure for each plot and combine the resulting plots using Bokeh layouts .

PDF generation, however, is more involved. Bokeh has functions to render documents as PNG and SVG, but there's not PDF generation yet. You can either submit a feature request via Github or implement something yourself. In the latter case, you can either generate PNG/SVG files and create PDF documents from them or get inspiration from bokeh.io.export_png and implement direct export to PDF using Selenium and PhantomJS.