adiar

adiar copied to clipboard

Build as a Shared Library

1

**Describe the bug** It is not possible to build Adiar as a shared library. **To Reproduce** Try to build the `make main` target changed such that CMake also is parsed...

SSoelvsten

Follow-up on #397 . In that refactor we provide a major clean-up of all the prior code. But, we can improve the code quality even more, by generalising the use...

SSoelvsten

Reduce: Use Adiar's I/O Wrappers

1

In the *reduce* algorithm, all mappings due to the first reduction rule are placed in a `tpie::file_stream`. With #397 we have cleaned up Adiar's wrappers on files and streams and...

SSoelvsten

Improve legibility of `bdd_support` (earlier `bdd_varprofile`)

1

It is not clear, that Adiar has support to obtain all variables in a BDD or ZDD. **Tasks**: - BDD: - [x] Move `bdd_varprofile` up next to `bdd_minvar` and `bdd_maxvar`...

SSoelvsten

Currently, the `max_1level_cut` and `max_2level_cut` functions for the `zdd` class adds an extra edge to account for edges being paired with the suppressed edge to *false*. Yet, specifically for the...

SSoelvsten

We should include a license statement in each of the files regardless of the license in question. ## Where to Include the License **Public files** Include the license in the...

SSoelvsten

Apply Second Reduction Rule inside of Builder

2

Right now, the `bdd_builder` and `zdd_builder` do not account for the user generating the same node twice. **Implement a Hash Table** - [ ] Implement a `levelized_hash_table` template that internally...

SSoelvsten

Dealing with Small BDDs using Depth-First Algorithms

1

With Adiar [v1.2](https://github.com/SSoelvsten/adiar/releases/tag/v1.2.0) we have been able to decrease the threshold as to when Adiar is viable in comparison to conventional depth-first BDD packages. Yet, this does not remove the...

SSoelvsten

The current over-approximations for 1-level and 2-level cuts within the `bdd_builder` and `zdd_builder` (specifically within the destructor of the `node_writer`) works well when the BDD does not have many far-going...

SSoelvsten

Discussion: Naming Convention for (Public) API

6

While looking at the public API as a whole, I have been wondering once again *"Is there really not a pretty interface?"*. I personally like *snake_case* and have a slight...

SSoelvsten