[MPI] Distribute text blocks across MPI processes

[hubertwojcik]

Goal

Implement MPI-based parallelism where rank 0 distributes text blocks to all processes and each runs an independent Aho-Corasick search.

Tasks

• Rank 0 reads the full text and scatters blocks to all processes using MPI_Scatterv
• Broadcast the pre-built automaton transition table to all ranks using MPI_Bcast
• Each rank searches its block and returns matches to rank 0 via MPI_Gatherv
• Rank 0 merges results, adjusting match positions by block offset
• Add CLI invocation via mpirun -np N aho_mpi --patterns FILE --text FILE

Acceptance Criteria

• Results match sequential baseline (excluding boundary patterns — handled in next issue)
• Runs correctly with 2, 4, and 8 processes
• No MPI errors or memory leaks detected with valgrind --tool=helgrind