Getting Started#

Use this page as an operational checklist for the four BCW repository examples, after the environment is ready and before you start reading the derivation-heavy walkthroughs.

The objective is simple: run the four scripts from the repository root and compare the output to a small set of headline magnitudes.

Step 1: Prepare The Repository Environment#

From the repository root:

uv sync
uv run python -c "import finhjb as fjb; print(fjb.__all__[:5])"

If you are on a headless machine, also set:

export MPLBACKEND=Agg

The BCW scripts are documented as repository-root examples. The supported usage pattern is:

uv run python src/example/BCW2011Liquidation.py

not local-directory execution from inside src/example/.

Step 2: Run Case I#

MPLBACKEND=Agg uv run python src/example/BCW2011Liquidation.py

Use w_bar ≈ 0.22, p'(0) ≈ 30, and strongly negative investment at very low cash as the first-pass checks.

Step 3: Run Case II#

MPLBACKEND=Agg uv run python src/example/BCW2011Refinancing.py

Use these magnitudes as the first checks:

with phi=1%: w_bar ≈ 0.19, m ≈ 0.06,
with phi=0: w_bar ≈ 0.14, m ≈ 0.

Step 4: Run Case IV#

MPLBACKEND=Agg uv run python src/example/BCW2011Hedging.py

For the costly-margin line, use w_- ≈ 0.07, w_+ ≈ 0.11, w_bar ≈ 0.14, and psi in [-5, 0] as the first checks.

Step 5: Run Case V#

MPLBACKEND=Agg uv run python src/example/BCW2011CreditLine.py

Use w_bar ≈ 0.08, c+m ≈ 0.10, and p'(0) near 1.01 when the credit line is active as the first checks.

Step 6: Read The Result Intentionally#

After a run, check these first:

print(bundle["artifact_paths"])
print(bundle["results"])

If you want the solved grid for one scenario:

result = bundle["results"]["fixed-cost"]
print(result["summary"])
print(result["grid"].df.head())
print(result["grid"].df.tail())