These are the files that represent the only changes the bot would have actually committed if let loose 'in the wild,' as the vast majority of Bulgarian entries near the top of the category already have definitions of some kind. These are the ones without any existing Bulgarian definition. Hopefully there are no problems to see here.
These are the text files containing the proposed wikitext that the bot would have saved on the corresponding pages for each file. Inside is what would be the code for the file (but ignoring the Bulgarian-entry check, i.e. attempting to place new derived form entries regardless of whether there already existed one or not.
This file should serve as an example for what the generated dictionary 'pages_to_create' will look like. This example is naturally in JSON, but should convey the same information as what occurs in the Python script: a dictionary with the structure seen herein is generated and subsequently used to iterate over a series of pages, providing exactly the data required to generate the contents of each page. (The example seen here is the result of harvesting the declined-form data from https://en.wiktionary.org/wiki/%D0%BA%D1%83%D0%BA%D0%BB%D0%B0.) Each page title within the dictionary contains an array under the key "associations", which itself contains a number of objects with the following contents: one array "mapping", which contains tuples of (lemma, derived form) pairs. In other words, this identifies, for each title, one derived form, with reference to the original lemma. (As there can be multiple etymologies for a derived term, by the way, this is why there are multiple lemma and derived-form fields: the stress, for example, could differ between different etymologies. In this example, they incidentally do not.); the other content being an array "forms", which contains tuples of (form, number) pairs - in other words, specifying the data of what specific type of derived form this is. The exception is when the form is the 'count form', which doesn't exist in the singular and is hence given a special designation when used in the derived form template: whereas e.g. definite singular maps simply to {{...def|s}}, the count form doesn't require a plural/singular distinction, since it's always plural; its designation is {{...count|form}}.
The dictionary generated in the article creation process is used as described above to add content to pages. Please do inspect the function of the script yourself if you are interested or would like to expound any errors. Thanks for reading!
Expanded functionality of the bot to create derived forms for multiple etymologies at a time; created a loop to iterate over all lemmas in the category for "Bulgarian nouns". Probably various other improvements that I cannot recall, however.
