Add tracing spans and timing trees to the CLI

[zkfriendly]

The CLI now prints nested proof stages and durations to stderr while preserving the existing stdout summary.
Covers setup, witness generation, Spartan, BitZ, GKR, PCS, and verification.
It supports RUST_LOG filtering, terminal colors, NO_COLOR, and --quiet for benchmark runs.

Sample Output:

╮ bitz_cli::cmd::circuit_e2e::circuit_e2e circuit=sha256-chain threads=16 field="Q100" pcs="Fast" hash="Blake3"
├─╮ bitz_cli::cmd::circuit_e2e::generate_inputs
├─╯ generate_inputs: 84.75µs duration
├─╮ bitz_cli::end_to_end::setup
├─╯ setup: 133.60ms duration
│ 134.15ms INFO Circuit prepared opening_path=Virtual constraints=5824 assignment_bits=168001 committed_bits=55104 padded_committed_bits=4194304
├─╮ bitz_cli::end_to_end::witness
├─╯ witness: 1.11ms duration
├─╮ bitz_cli::end_to_end::commit
│ ├─╮ pcs::commitment::Commit witness
│ ├─╯ Commit witness: 760.33µs duration
├─╯ commit: 769.04µs duration
├─╮ bitz_cli::end_to_end::prove opening_path=Virtual
│ ├─╮ spartan::piop::Prove Spartan
│ │ ├─╮ spartan::sumcheck::Prove outer sumcheck
│ │ ├─╯ Prove outer sumcheck: 727.54µs duration
│ │ ├─╮ spartan::matrix::Bind Spartan matrices
│ │ ├─╯ Bind Spartan matrices: 1.50ms duration
│ │ ├─╮ spartan::sumcheck::Prove inner sumcheck
│ │ ├─╯ Prove inner sumcheck: 1.96ms duration
│ ├─╯ Prove Spartan: 4.34ms duration

Validation:

• cargo test --release -p bitz-cli: all 14 tests passed, including four span formatter tests.
• cargo clippy --workspace --all-targets
• cargo fmt --all --check
• Checked proof output, quiet mode, filters, colors, and error output through the CLI.

[shreyas-londhe]

LGTM, thanks for the PR!