From 7c73974cd476a297baee71519e04e9b466d20b86 Mon Sep 17 00:00:00 2001 From: Marine Denolle Date: Tue, 7 Jul 2026 14:59:33 -0700 Subject: [PATCH] Add the workflow chart to the narrative site (navbar "Workflow") Publish the codameter workflow chart (pipeline + agent steps + evaluation breakdown) as a standalone page on the Quarto site and link it from the navbar. - quarto/workflow.html: self-contained page (own design, back-to-site bar), kept as raw HTML so its CSS does not fight the theme. - quarto/_quarto.yml: register it as a project resource and add a top-level "Workflow" navbar item (href: workflow.html). - quarto/_site: re-rendered (freeze restored, no re-execution) so every page carries the new navbar link and _site/workflow.html ships. The deploy workflow publishes quarto/_site to gh-pages, so this goes live on the next master/main deploy. Co-Authored-By: Claude Opus 4.8 (1M context) --- quarto/_quarto.yml | 6 + quarto/_site/index.html | 4 + quarto/_site/survey-best-practices.html | 4 + quarto/_site/survey-references.html | 4 + quarto/_site/survey-synthetic-demo.html | 4 + quarto/_site/theory-bayesian-measurement.html | 4 + quarto/_site/theory-measurement-uq.html | 4 + quarto/_site/theory-processing-depth.html | 4 + quarto/_site/theory-uq.html | 4 + quarto/_site/tutorial-00-measurement.html | 4 + quarto/_site/tutorial-00-measurement.ipynb | 14 +- quarto/_site/tutorial-01-data.html | 4 + quarto/_site/tutorial-02-forward.html | 4 + quarto/_site/tutorial-03-inversion.html | 4 + quarto/_site/tutorial-04-coupling.html | 4 + quarto/_site/tutorial-05-interpretation.html | 4 + quarto/_site/tutorial-06-endtoend.html | 4 + quarto/_site/uncertainty-3-stress.html | 4 + quarto/_site/uncertainty-4-attribution.html | 4 + quarto/_site/workflow.html | 437 ++++++++++++++++++ quarto/workflow.html | 437 ++++++++++++++++++ 21 files changed, 955 insertions(+), 7 deletions(-) create mode 100644 quarto/_site/workflow.html create mode 100644 quarto/workflow.html diff --git a/quarto/_quarto.yml b/quarto/_quarto.yml index 9e03fe6..387a785 100644 --- a/quarto/_quarto.yml +++ b/quarto/_quarto.yml @@ -3,6 +3,10 @@ project: output-dir: _site render: - "*.qmd" + resources: + # Standalone workflow chart (self-contained design; served verbatim and + # linked from the navbar). Kept as raw HTML so its CSS never fights the theme. + - workflow.html execute: freeze: auto # cache executed cells; re-run only when the .qmd changes @@ -25,6 +29,8 @@ website: left: - text: "Home" file: index.qmd + - text: "Workflow" + href: workflow.html - text: "Theory" menu: - text: "1 · Measurement — marginal errors" diff --git a/quarto/_site/index.html b/quarto/_site/index.html index 84263f7..264cde6 100644 --- a/quarto/_site/index.html +++ b/quarto/_site/index.html @@ -329,6 +329,10 @@
  • Home +
    • +
  • + +Workflow
  • Home +
    • +
  • + +Workflow
  • Home +
    • +
  • + +Workflow
  • Home +
    • +
  • + +Workflow
  • Home +
    • +
  • + +Workflow
  • Home +
    • +
  • + +Workflow
  • Home +
    • +
  • + +Workflow
  • Home +
    • +
  • + +Workflow
  • Home +
    • +
  • + +Workflow
  • Home +
    • +
  • + +Workflow
  • Home +
    • +
  • + +Workflow
  • Home +
    • +
  • + +Workflow
  • Home +
    • +
  • + +Workflow
  • Home +
    • +
  • + +Workflow
  • Home +
    • +
  • + +Workflow
  • Home +
    • +
  • + +Workflow
  • Home +
    • +
  • + +Workflow
  • ← codameter/ workflowGitHub
    +
    +
    +

    codameter · ambient-noise dv/v monitoring

    +

    From raw coda to stress at depth — and where agents fit

    +

    codameter turns cross-correlation functions into a relative velocity-change + (δv/v) time series, propagates its uncertainty, and inverts it for depth and stress. + This chart maps that pipeline, marks the steps an agent can drive, and decomposes the + evaluation loop into its smallest reproducible parts.

    +
    + Agent-amenable — an LLM can do or drive it + Deterministic compute — pure code, reproducible + Data artifact — a dataset or output object + Human-in-the-loop — needs a scientist's call +
    +
    + +
    +
    + design rule +

    Agents propose; deterministic oracles score. Every agent step is bounded by + code that regenerates a known truth and measures the result — so trust never rests on the model.

    +
    +
    + + +
    +

    A · the scientific pipeline

    +

    Measurement to interpretation

    +

    The core dv/v arc. The processing at stage 2 is a long chain of choices; the whole + point of the advisor and the golden set is to make those choices explicit and testable.

    +
    +
    +
    +
    compute + human QC
    +

    1. Ingest & QC

    +

    Raw noise / CCFs, gap detection, time alignment, SNR gating.

    + data.loaders · qc · readiness (Phase 0) +
    + → +
    +
    compute · choice=agent
    +

    2. Measure δv/v

    +

    Estimator, band, coda window, reference, stacking, aggregation → δv/v(t).

    + synthetic_demo · deviations.run_pipeline +
    + → +
    +
    compute
    +

    3. Uncertainty → C d

    +

    Aleatoric floor + processing multiverse, marginalized into one covariance.

    + uq_measurement · uq_processing · uq_bayes +
    + → +
    +
    compute
    +

    4. Depth inversion

    +

    Frequency-dependent kernels → a shear-velocity change vs depth.

    + uq_depth · kernels (disba) +
    + → +
    +
    compute + human
    +

    5. Stress & strain

    +

    Petrophysical β-bridge from velocity change to stress at depth.

    + interpretation · forward · coupling +
    + → +
    +
    compute · draft=agent
    +

    6. Anomaly & attribution

    +

    Detect residual transients; attribute to a forcing mechanism.

    + anomaly · forcing_models +
    +
    +
    +

    Agent openings: stage 2's choices come from the advisor; stages 5–6 + benefit from agent-drafted interpretation, but the numbers stay code-owned.

    +
    + + +
    +

    B · the parameter advisor

    +

    Elicit → map → validate → report

    +

    The codameter-advisor + skill front-ends stage 2. It never invents parameters: it maps a use case to a config, then proves + it on a matched synthetic before recommending.

    +
    +
    +
    +
    agent + human
    +

    1. Elicit use case

    +

    Ask the monitoring target, band, geometry, expected amplitude, data problems.

    + use_cases.ELICITATION · AskUserQuestion +
    + → +
    +
    compute · surfaced by agent
    +

    2. Map to config

    +

    Look up the recommended choice set with cited rationale.

    + use_cases.recommend() +
    + → +
    +
    compute · agent-invoked
    +

    3. Validate live

    +

    Run recommended vs naive on a matched golden case; report the RMS cost.

    + golden.generate · run_pipeline · multiverse +
    + → +
    +
    agent
    +

    4. Report

    +

    Config + YAML + numbers + the one caveat that matters for this case.

    + references/report_format.md +
    + ↶ +
    +
    +

    Loop: if the user pins an axis, re-map and re-validate. The report carries the + measured contrast, not an opinion.

    +
    + + +
    +

    C · evaluation, decomposed

    +

    Two lanes of small, checkable steps

    +

    "Evaluation" is really two pipelines. The left lane scores a pipeline against a + known truth (the regression oracle). The right lane scores an agent against that same oracle + (the FrugalMind benchmark). They share their scoring core, so an agent is judged by the same code that + guards the pipeline.

    +
    + +
    +

    Lane 1 · Pipeline scoring (deterministic oracle)

    +

    Reproducible from seeds. Runs in CI and fans out on Fargate.

    +
      +
    1. + 1 +

      Define the case human / agent

      +

      Pick a use case and regime (mainstream or an edge: low-SNR, clock drift, freq-dependent, sparse).

      + golden.CASES +
      2. +
    2. + 2 +

      Generate golden synthetic data

      +

      Seed → daily CCFs with an exactly known δv/v(t). Deterministic; cached, hash-keyed.

      + golden.generate() +
      3. +
    3. + 3 +

      Produce a config agent or code

      +

      The recommended config, or a candidate to test. This is the only lever under study.

      + use_cases.recommend() +
      4. +
    4. + 4 +

      Run the pipeline compute

      +

      Recover δv/v(t) under that config, with the estimator, reference and stacking it names.

      + deviations.run_pipeline() +
      5. +
    5. + 5 +

      Score recovery compute

      +

      Baseline-aligned RMS of recovered vs true δv/v (the DC offset is unobservable, so removed).

      + golden._rms() +
      6. +
    6. + 6 +

      Sweep & shard compute · scale

      +

      Every case × every config in a grid; round-robin shards across an array of tasks.

      + bench.sweep · --shard k/N +
      7. +
    7. + 7 +

      Aggregate & drift-guard data

      +

      Merge shards, pick best config per case, and lock expected RMS in the committed manifest.

      + bench.aggregate · manifest.json +
      8. +
    +
    + +
    +

    Lane 2 · Agent scoring (FrugalMind benchmark)

    +

    Judges a model, cost-aware. Reuses Lane 1's steps 4–5 as its scorer.

    +
      +
    1. + 1 +

      Build rows data

      +

      Each case → a prompt + gold (case id + tolerances) + a scorer spec. No arrays leak.

      + frugalmind.build_rows() +
      2. +
    2. + 2 +

      Model answers agent · under test

      +

      Returns a config (JSON) for param_recommendation, + or the recovered series for dvv_series.

      + EvalRunner → adapter.generate() +
      3. +
    3. + 3 +

      Rebuild scorer compute

      +

      Reconstruct the deterministic scorer from the JSON spec — portable, no hidden state.

      + make_scorer_from_spec() +
      4. +
    4. + 4 +

      Score in [0,1] compute

      +

      Run the answer through Lane 1 (steps 4–5). Wrong band or cycle-skip → near 0; + a no-change series is null-anchored to 0.

      + dvv_recovery · dvv_series_regression +
      5. +
    5. + 5 +

      Route on budget compute

      +

      Track spend per model; stop cleanly at the cap. Cheapest model that clears the quality floor wins.

      + EvalRunner · BudgetGuard +
      6. +
    6. + 6 +

      Adversarial cross-check agent · optional

      +

      A panel re-derives δv/v with independent estimators; agreement gates the result.

      + multi-estimator verify (proposed) +
      7. +
    7. + 7 +

      Leaderboard data · human reads

      +

      Score × cost per model and suite; skill-lift over the no-skill baseline.

      + leaderboard · skill_lift +
      8. +
    +
    +
    +
    + + +
    +

    D · where agents plug in

    +

    Agent opportunities, ranked by leverage

    +

    The steps worth agentifying, what the agent actually does, and where each stands today. + Everything else stays deterministic on purpose.

    +
    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    StepAgent roleToolingStatus
    Use-case elicitationdrives asks & infers the monitoring settingcodameter-advisorbuilt
    Recommend & reportdrafts surfaces config + cited rationale + numbersadvisor · report_formatbuilt
    Config under testis tested picks a processing config for a scenarioparam_recommendationbuilt (benchmark)
    End-to-end seriesis tested runs codameter in a sandbox, returns δv/v(t)dvv_seriesbuilt; needs sandbox
    Interpretation draftassists narrates stress / anomaly attributioninterpretation · anomalycandidate
    Adversarial verifypanel cross-checks with independent estimatorsmulti-estimatorproposed
    Eval orchestrationmeta chooses sweeps, reads leaderboards, flags driftbench · frugalmindcandidate
    +
    +
    + +
    + Node type is encoded in color and label. Module names are the real codameter / frugalmind entry points. + Scoring is deterministic end to end; agents supply choices and prose, never the numbers. +
    +
    + + + diff --git a/quarto/workflow.html b/quarto/workflow.html new file mode 100644 index 0000000..88432f5 --- /dev/null +++ b/quarto/workflow.html @@ -0,0 +1,437 @@ + + + + + +Codameter workflow: pipeline, agent steps, evaluation breakdown + + + +
    ← codameter/ workflowGitHub
    +
    +
    +

    codameter · ambient-noise dv/v monitoring

    +

    From raw coda to stress at depth — and where agents fit

    +

    codameter turns cross-correlation functions into a relative velocity-change + (δv/v) time series, propagates its uncertainty, and inverts it for depth and stress. + This chart maps that pipeline, marks the steps an agent can drive, and decomposes the + evaluation loop into its smallest reproducible parts.

    +
    + Agent-amenable — an LLM can do or drive it + Deterministic compute — pure code, reproducible + Data artifact — a dataset or output object + Human-in-the-loop — needs a scientist's call +
    +
    + +
    +
    + design rule +

    Agents propose; deterministic oracles score. Every agent step is bounded by + code that regenerates a known truth and measures the result — so trust never rests on the model.

    +
    +
    + + +
    +

    A · the scientific pipeline

    +

    Measurement to interpretation

    +

    The core dv/v arc. The processing at stage 2 is a long chain of choices; the whole + point of the advisor and the golden set is to make those choices explicit and testable.

    +
    +
    +
    +
    compute + human QC
    +

    1. Ingest & QC

    +

    Raw noise / CCFs, gap detection, time alignment, SNR gating.

    + data.loaders · qc · readiness (Phase 0) +
    + → +
    +
    compute · choice=agent
    +

    2. Measure δv/v

    +

    Estimator, band, coda window, reference, stacking, aggregation → δv/v(t).

    + synthetic_demo · deviations.run_pipeline +
    + → +
    +
    compute
    +

    3. Uncertainty → C d

    +

    Aleatoric floor + processing multiverse, marginalized into one covariance.

    + uq_measurement · uq_processing · uq_bayes +
    + → +
    +
    compute
    +

    4. Depth inversion

    +

    Frequency-dependent kernels → a shear-velocity change vs depth.

    + uq_depth · kernels (disba) +
    + → +
    +
    compute + human
    +

    5. Stress & strain

    +

    Petrophysical β-bridge from velocity change to stress at depth.

    + interpretation · forward · coupling +
    + → +
    +
    compute · draft=agent
    +

    6. Anomaly & attribution

    +

    Detect residual transients; attribute to a forcing mechanism.

    + anomaly · forcing_models +
    +
    +
    +

    Agent openings: stage 2's choices come from the advisor; stages 5–6 + benefit from agent-drafted interpretation, but the numbers stay code-owned.

    +
    + + +
    +

    B · the parameter advisor

    +

    Elicit → map → validate → report

    +

    The codameter-advisor + skill front-ends stage 2. It never invents parameters: it maps a use case to a config, then proves + it on a matched synthetic before recommending.

    +
    +
    +
    +
    agent + human
    +

    1. Elicit use case

    +

    Ask the monitoring target, band, geometry, expected amplitude, data problems.

    + use_cases.ELICITATION · AskUserQuestion +
    + → +
    +
    compute · surfaced by agent
    +

    2. Map to config

    +

    Look up the recommended choice set with cited rationale.

    + use_cases.recommend() +
    + → +
    +
    compute · agent-invoked
    +

    3. Validate live

    +

    Run recommended vs naive on a matched golden case; report the RMS cost.

    + golden.generate · run_pipeline · multiverse +
    + → +
    +
    agent
    +

    4. Report

    +

    Config + YAML + numbers + the one caveat that matters for this case.

    + references/report_format.md +
    + ↶ +
    +
    +

    Loop: if the user pins an axis, re-map and re-validate. The report carries the + measured contrast, not an opinion.

    +
    + + +
    +

    C · evaluation, decomposed

    +

    Two lanes of small, checkable steps

    +

    "Evaluation" is really two pipelines. The left lane scores a pipeline against a + known truth (the regression oracle). The right lane scores an agent against that same oracle + (the FrugalMind benchmark). They share their scoring core, so an agent is judged by the same code that + guards the pipeline.

    +
    + +
    +

    Lane 1 · Pipeline scoring (deterministic oracle)

    +

    Reproducible from seeds. Runs in CI and fans out on Fargate.

    +
      +
    1. + 1 +

      Define the case human / agent

      +

      Pick a use case and regime (mainstream or an edge: low-SNR, clock drift, freq-dependent, sparse).

      + golden.CASES +
      2. +
    2. + 2 +

      Generate golden synthetic data

      +

      Seed → daily CCFs with an exactly known δv/v(t). Deterministic; cached, hash-keyed.

      + golden.generate() +
      3. +
    3. + 3 +

      Produce a config agent or code

      +

      The recommended config, or a candidate to test. This is the only lever under study.

      + use_cases.recommend() +
      4. +
    4. + 4 +

      Run the pipeline compute

      +

      Recover δv/v(t) under that config, with the estimator, reference and stacking it names.

      + deviations.run_pipeline() +
      5. +
    5. + 5 +

      Score recovery compute

      +

      Baseline-aligned RMS of recovered vs true δv/v (the DC offset is unobservable, so removed).

      + golden._rms() +
      6. +
    6. + 6 +

      Sweep & shard compute · scale

      +

      Every case × every config in a grid; round-robin shards across an array of tasks.

      + bench.sweep · --shard k/N +
      7. +
    7. + 7 +

      Aggregate & drift-guard data

      +

      Merge shards, pick best config per case, and lock expected RMS in the committed manifest.

      + bench.aggregate · manifest.json +
      8. +
    +
    + +
    +

    Lane 2 · Agent scoring (FrugalMind benchmark)

    +

    Judges a model, cost-aware. Reuses Lane 1's steps 4–5 as its scorer.

    +
      +
    1. + 1 +

      Build rows data

      +

      Each case → a prompt + gold (case id + tolerances) + a scorer spec. No arrays leak.

      + frugalmind.build_rows() +
      2. +
    2. + 2 +

      Model answers agent · under test

      +

      Returns a config (JSON) for param_recommendation, + or the recovered series for dvv_series.

      + EvalRunner → adapter.generate() +
      3. +
    3. + 3 +

      Rebuild scorer compute

      +

      Reconstruct the deterministic scorer from the JSON spec — portable, no hidden state.

      + make_scorer_from_spec() +
      4. +
    4. + 4 +

      Score in [0,1] compute

      +

      Run the answer through Lane 1 (steps 4–5). Wrong band or cycle-skip → near 0; + a no-change series is null-anchored to 0.

      + dvv_recovery · dvv_series_regression +
      5. +
    5. + 5 +

      Route on budget compute

      +

      Track spend per model; stop cleanly at the cap. Cheapest model that clears the quality floor wins.

      + EvalRunner · BudgetGuard +
      6. +
    6. + 6 +

      Adversarial cross-check agent · optional

      +

      A panel re-derives δv/v with independent estimators; agreement gates the result.

      + multi-estimator verify (proposed) +
      7. +
    7. + 7 +

      Leaderboard data · human reads

      +

      Score × cost per model and suite; skill-lift over the no-skill baseline.

      + leaderboard · skill_lift +
      8. +
    +
    +
    +
    + + +
    +

    D · where agents plug in

    +

    Agent opportunities, ranked by leverage

    +

    The steps worth agentifying, what the agent actually does, and where each stands today. + Everything else stays deterministic on purpose.

    +
    + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + +
    StepAgent roleToolingStatus
    Use-case elicitationdrives asks & infers the monitoring settingcodameter-advisorbuilt
    Recommend & reportdrafts surfaces config + cited rationale + numbersadvisor · report_formatbuilt
    Config under testis tested picks a processing config for a scenarioparam_recommendationbuilt (benchmark)
    End-to-end seriesis tested runs codameter in a sandbox, returns δv/v(t)dvv_seriesbuilt; needs sandbox
    Interpretation draftassists narrates stress / anomaly attributioninterpretation · anomalycandidate
    Adversarial verifypanel cross-checks with independent estimatorsmulti-estimatorproposed
    Eval orchestrationmeta chooses sweeps, reads leaderboards, flags driftbench · frugalmindcandidate
    +
    +
    + + +
    + + +