Request for LYR, TTM, LF computation for PIT data

🌟 Feature Description

When using pit data for computing, there are usually three ways to deal with them, which are lf, lyr and ttm. lf processing means the last file, which returns the value from last financial report; lyr means the last year ratio, which returns the value from the latest yearly report; ttm means trailing 12 months, which returns the mean value of the last 4 quarters.

Motivation

1. Scenario: When using pit data for caculation some indicator, the processing is neccessary, for the quarterly data may increase with the quarter. In order to solve the problem, we usually use the latest 4 quarters mean value instead. However, the pit database store only the quarterly data, we need a method to transform them into other forms.
2. References: Like ricequant, there is a suffix like "_mrq_0", "_lf", "lyr" or "_ttm" to specify the type of data you want to fetch. See here for more information.

Alternatives

In fact, we can store pit data as a single feature in the directory of daily frequency stock data. So do the "ttm", "lyr" and "lf" data. However, storing in this way cannot maximize the potential of the pit database. So maybe developping a new file suffix like _ttm instead of just choosing between _a and _q can help a lot.

Or if all the data can be computed dynamically, and only raw quarterly data saved in our database is perfect. I’m thinking about developing operators like TTM, LF specifically applied on pit data, and maybe this is the best solution to the problem

No files, tests, or entry points are named. Start by locating the existing PIT data processing and operator paths, then clarify how LF, LYR, and TTM should be represented and calculated. Done means the chosen approach supports these transformations from quarterly PIT data with documented behavior and coverage.