ST partitions things, the same IC's are used on the Nucleo boards, are used everywhere else. The functionality of the chips is described in the Reference Manuals (RM), the Data Manual (DM) cover performance at pin/device level where the bond-out of the same IC is constrained by the pin count of the different package. Only a subset of functionality can be escaped on any specific device.
The User Manuals for the assorted Nucleo boards typically cover a whole group of related board designs. These provide pin maps and details of solder bridges, default, LEDs and UARTs are covered here. The UM used to have schematics, but as the boards have expanded and changed over the years the schematics are now separate files, usually in "CAD Resources" rather than "Documentation" tabs. One board is designed to support multiple chips, and the default options allow for the most common usage, whilst facilitating alternate usage. This often impacts the usage of Arduino shields, how those might be wired, and how limited resources like UART might be wired. See NUCLEO-64 typically share Arduino D0/D1 pins with those used by the ST-LINK's VCP UART.
The schematics are a good point of reference, there are also Include files as part of the CubeF4 packages for various boards, via BSP Board Support Package.
Historically one might use AWK, or other scripting, to process .SVD files or Compiler Pre-Processor output to extract defines, structures and details in export them in a form suitable for assemblers. ie use EQU, create .INC files, etc. Modern assemblers can frequently just pull include files.
The market tends to focus on what the bulk of customers are using, so the compilers and include files will get all the attention.
Between the DM and RM you should be able to get the base addresses of all the peripherals, and register offsets within each of these instances. For the processor cores either the ARM TRM or ST Programming Manuals (PM) should cover the Cortex-M, along with resources like Joseph Yiu's books on the more practical side.
