Nfa compose #389
Nfa compose #389
Conversation
…insert_word methods
…en inserting a word.
There was a problem hiding this comment.
The PR looks good to me in general. The plan now is to merge this PR immediately, so I can apply the changes here on string solving. Resolve the review issues later on, when there is time.
Furthermore, the things that should be modifed are:
- nfa::Nfa::insert_word() can have optional target word
- nft::Nft::insert_word() can have optional target word
- nft::Nft::insert_word_by_parts() can have optional target word
- nft::Nft::insert_identity() has an overload with a vector of symbols, creating an identity for all the symbols within.
| * @param word The nonempty word to be inserted into the NFA. | ||
| * @param tgt The target state where the word ends. It must already be a part of the automaton. | ||
| */ | ||
| void insert_word(const State src, const Word &word, const State tgt); |
There was a problem hiding this comment.
Future TODO: The target state should be optional and be returned from the function afterwards.
There was a problem hiding this comment.
So, when the target state will not be specified, the method will create a sequence of states from src (already existing) to tgt (newly created), where tgt does not have any successors and is not a final state. It that correct?
There was a problem hiding this comment.
Yes. And the newly created target will be returned from the function so that the user knows where to continue.
| * @param[in] lhs_first_aux_state The first auxiliary state in @p lhs. Two auxiliary states can not form a product state. | ||
| * @param[in] rhs_first_aux_state The first auxiliary state in @p rhs. Two auxiliary states con not form a product state. | ||
| * @return NFT as a product of NFTs @p lhs and @p rhs with ε handled as regular symbols. |
There was a problem hiding this comment.
I would explain here what are these auxiliary states for.
| */ | ||
| Nft product(const Nft& lhs, const Nft& rhs, const std::function<bool(State,State)> && final_condition, | ||
| const Symbol first_epsilon = EPSILON, std::unordered_map<std::pair<State,State>, State> *prod_map = nullptr); | ||
| std::unordered_map<std::pair<State,State>, State> *prod_map = nullptr, | ||
| const State lhs_first_aux_state = Limits::max_state, const State rhs_first_aux_state = Limits::max_state); |
There was a problem hiding this comment.
Are the states used anywhere in Delta? Because if so, the Limits::max_state of memory is being allocated.
There was a problem hiding this comment.
No, these states are not used in Delta if they are not already a part of it. Essentially, it prevents two states from Delta, which are greater than the specified auxiliary state (implying that these states are auxiliary), from forming a product state.
There was a problem hiding this comment.
You are right. I kind of saw it in the code later on, but I wanted to make sure.
|
|
||
| #include <fstream> |
There was a problem hiding this comment.
Unnecessary import, probably left behind from debugging.
This pull request includes:
nft::project_out().Nft::insert_word(),Nft::insert_word_by_parts(), andNft::insert_identity()for inserting a given word (as one word or a vector of its parts, each for a different level) between source and target states (source and target states are equal for the identity).nft::insert_level()andnft::insert_levels()for inserting levels with adefault_symbolon transitions with newly inserted levels. These functions are utilized in matching levels of synchronization between two transducers.nft::intersection()andnft::algorithm::product()to handle transducers withEPSILONandDONT_CAREtransitions and a set of auxiliary states (two auxiliary states cannot create a product state).nft::composefor computing the composition of two transducers.