run_st_job

Overview

run_st_job builds a stress-testing HTML report (vine copula, Gaussian, or Student-t shocks). Inputs:

portfolio_optimization_*.json — usual path
data_extractor_*.json — fast path; optional portfolio_weights, else equal 1/N

Describe the shock in chat (e.g. “AAPL −15%, Student-t”); skills set stress_dict and mode.

Parameters

string

required

Blob URL to a portfolio_optimization_*.json or data_extractor_*.json artifact.

object

required

Stress test configuration.

Show common config fields

string

required

Stress model. Supported values: vine_copula, gaussian, student_t (alias: student).

object

required

Required ticker-to-shock mapping, for example { "AAPL": -0.05 }. For gaussian / student_t, you must leave at least one asset unstressed (do not put every portfolio ticker in stress_dict).

number

Number of simulations. Example: 5000.

object

Optional weights for the data-extraction fast path, e.g. { "AAPL": 0.5, "MSFT": 0.5 }. If omitted on that path, equal 1/N weights are used.

number

student_t only. Default: 4; must be greater than 2. Lower values mean heavier tails.

array

Optional list of tickers — for example ["SPY", "MTUM"] — added to the price universe with zero weight, so their mean impacts appear in the result table alongside your holdings. Works in all three modes, and sits at the top level of the config like everything else here.

array

vine_copula only. Tickers to place at the centre of the vine structure.

string

vine_copula only. "kendall_tau", "spearman" or "pearson".

boolean

vine_copula only. Pick the central assets by clustering instead of naming them.

integer

Number of clusters, used together with use_cluster_analysis.

The config is flat: every field above sits at the top level. There is no nested block per mode, and factors in particular is not written the way run_risk_job writes its factor_impact parameters.

Returns

{
  "status": "Succeeded",
  "output_url": "https://stqsnpprod.blob.core.windows.net/data/stress_test_gaussian_e2f3a4b5c6d7.html",
  "output_name": "stress_test_gaussian_e2f3a4b5c6d7.html",
  "execution_name": "job-st-worker-abc123xyz"
}

Example

{
  "input_url": "https://stqsnpprod.blob.core.windows.net/data/portfolio_optimization_35c1b5239d23.json",
  "config": {
    "stress_test_mode": "gaussian",
    "stress_dict": {
      "AAPL": -0.05,
      "MSFT": -0.03
    },
    "n_samples": 5000
  }
}

Next Step

The report is self-contained — no other tool reads a stress_test_ file. From here you would normally either rerun the same shock under a different stress_test_mode, since Gaussian, Student-t and vine copula disagree by construction, or look at the same book through run_risk_job.

Overview

How To

Tools description

Overview

Parameters

Returns

Example

Next Step

​Overview

​Parameters

​Returns

​Example

​Next Step

Overview

Parameters

Returns

Example

Next Step