These sites are built using a WYSIWYG editor. The pages have an unreliable, ad-hoc structure, and each page usually contains multiple recipes (or recipe variants). The number of servings is rarely given.
Hence, we have to ask the user to choose which ingredient/method lists are wanted (using completing-read-multiple), and concatenate them together. (Alternatively we could change the org-chef format to allow for labelled sections within the ingredients and method.)
Due to the unreliable page structure, my code searches the DOM for <ul> and <ol> elements to find recipes. This seems like a relatively robust strategy.
As a side note, a similar approach might be used to make a general fallback parser, since I assume most sites use <ul> for ingredients and <ol> for instructions.
See #56 and #15.
These sites are built using a WYSIWYG editor. The pages have an unreliable, ad-hoc structure, and each page usually contains multiple recipes (or recipe variants). The number of servings is rarely given.
Hence, we have to ask the user to choose which ingredient/method lists are wanted (using
completing-read-multiple
), and concatenate them together. (Alternatively we could change theorg-chef
format to allow for labelled sections within the ingredients and method.)Due to the unreliable page structure, my code searches the DOM for
<ul>
and<ol>
elements to find recipes. This seems like a relatively robust strategy.As a side note, a similar approach might be used to make a general fallback parser, since I assume most sites use
<ul>
for ingredients and<ol>
for instructions.