How Sefaria Treats Commentaries

In the Sefaria database, commentaries are treated the same as regular texts. This means that commentary texts have their own individual Index and each edition of text for that for that commentary is represented by its own Version record.

Index Attributes for Commentaries

We use specific attributes on the Index record to indicate that an Index is a commentary on another book. Those attributes are elaborated below:

Attribute	Type	Description
`dependence`	`String`	Generally `Commentary` or `Targum`. This field is used to denote commentaries and other potential non-standalone texts.
`base_text_titles`	`List` of `Strings`	The base book(s) this one is dependent on. For example, `Rashi on Genesis` is dependent on `Genesis`.
`base_text_mapping`	`String`	Enables auto-linking between a base text and a commentary. These values match keys in the `class_map` found in sefaria.helper.link.AutoLinkerFactory. So far, four options exist: `"one_to_one"`, `"many_to_one"`, `"many_to_one_default_only"` and `one_to_one_default_only`. `many_to_one` - The typical case of multiple commentary comments on a single segment of the base text. For example, there are many Rashi comments per verse of Tanakh. Tanakh is a depth-2 text, and Rashi on Tanakh is depth-3. `one_to_one` - Only used for commentaries defined as `targum`. They are linked to their original texts one-to-one since they are structured exactly the same as their base text - both Targum on Amos and Amos are depth-2 (a second example would be Steinsaltz on Bavli, the commentary has the same depth as the text, and also functions as a `one-to-one` commentary). The corresponding `default` keys work the same, except `default` nodes will be linked and other nodes will be ignored. (This helps maintain the auto-linking feature once afforded commentaries by default. It also makes room to expand and have various algorithms for linking between a dependent text and its base text/s)
`collective_title`	`String`	This is the value for a group of index records. It is used to contain the former commentator name which was once the primary handle on a commentary. Now used to group many indices belonging to a larger corpus, such as commentaries. Requires a matching Hebrew `Term`. ex: The `collective_title` for `Rashi on Genesis` is the same as the `Index` title, `Rashi`.
`is_cited`	`Boolean`	Only texts with `is_cited` set to `True` will be picked up as a citation in referencing texts. Default is `True`.

Below is an example of how these specific fields would be filled on the Index record of Ramban on Genesis:

"title": "Ramban on Genesis"
"base_text_titles": [
   "Genesis"
],
"base_text_mapping": "many_to_one",
"collective_title": "Ramban",
"dependence": "Commentary",
"categories": ["Tanakh", "Rishonim on Tanakh", "Ramban", "Torah"]

Commentaries and Category

Categories are mostly used to denote where the book will show up in the Sefaria Table of Contents.
This means that the text "Ramban on Genesis" with the category path of ["Tanakh", "Rishonim on Tanakh", "Ramban", "Torah"]will show up under Tanakh >> Rishonim on Tanakh >> Ramban >> Torah.

For a commentary, the category indicating the type of commentary (i.e. in this example, Rishonim on Tanakh) come after one top level category . E.g.["Tanakh", "Rishonim on Tanakh", "Ramban", "Torah"]
In some cases, it comes after multiple categories, particularly for Talmud. E.g.["Talmud", "Bavli", "Rishonim on Talmud", "Rashi", "Seder Zeraim" ]