pycsamt.agents.tipper_analysis#

pycsamt.agents.tipper_analysis#

TipperAnalysisAgent — Dedicated tipper vector analysis.

Computes induction arrows (Wiese or Parkinson convention), tipper magnitude and phase curves, and produces a map-view arrow plot alongside per-station tipper vs period profiles. Tipper vectors are a primary diagnostic for 3-D structure, crustal conductors, and coastal effects.

Wraps#

  • induction_arrows()

  • tipper_amp_phase()

  • Tipper

Classes

TipperAnalysisAgent(*[, api_key, model, ...])

Analyse tipper vectors and plot induction arrows.
class pycsamt.agents.tipper_analysis.TipperAnalysisAgent(*, api_key=None, model=None, llm_provider='claude', convention='wiese', use_imag=False, period_ref=None)[source]#

Bases: BaseAgent

Analyse tipper vectors and plot induction arrows.

Parameters:

  • api_key (str)

  • model (str)

  • llm_provider (str)

  • convention (str, optional) – Arrow convention. Wiese (default) points toward conductors in the real-part convention; Parkinson points toward them.

  • use_imag (bool, optional) – Use imaginary tipper parts for deeper structure (default False).

  • period_ref (arrow_table pandas.DataFrame — induction arrows at) – Reference period (s) for the induction arrow map. None uses the geometric mean of available periods.

  • keys (Output data)

  • ----------

  • path (sites /)

  • convention

  • use_imag

  • period_ref

  • output_dir (str, optional)

  • keys

  • ----------------

  • per-(station (tipper_table pandas.DataFrame —)

  • period)

  • period_ref

  • map (period_ref float — period used for arrow)

  • int (n_stations_with_tipper)

  • dict (figure_paths)

  • dict

Examples

>>> agent = TipperAnalysisAgent(convention='wiese')
>>> r = agent.execute({"path": "/data/WILLY_EDIs"})
>>> r["n_stations_with_tipper"]
12
SYSTEM_PROMPT: str = 'You are an expert in magnetotelluric tipper analysis and 3-D structure interpretation.\nGiven a tipper analysis result, write 4-5 sentences that:\n1. Describe the general tipper magnitude pattern across the survey (strong/weak, frequency dependence).\n2. Identify stations with anomalously large tipper amplitudes and their likely cause.\n3. Interpret the induction arrow direction(s) — do they point toward or away from a conductor?\n4. Assess the 3-D character of the survey based on tipper consistency along the profile.\n5. Recommend whether a 3-D inversion is warranted or whether 2-D is sufficient.\nReply in plain scientific English.\n'#

Override in subclasses to give the LLM its domain expertise.

execute(input_data)[source]#

Run this agent on input_data and return an AgentResult.

Subclasses must implement this method. The contract:

  • Reset self._last_cost = 0.0 at the top.

  • Record wall-clock time with t0 = time.time().

  • Return AgentResult(elapsed_seconds=time.time()-t0, cost_estimate_usd=self._last_cost, ...).

Parameters:

input_data (dict[str, Any])

Return type:

AgentResult