simudyne-pulse 0.7.0.dev6__tar.gz → 0.7.2__tar.gz

{simudyne_pulse-0.7.0.dev6/src/simudyne_pulse.egg-info → simudyne_pulse-0.7.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: simudyne-pulse
-Version: 0.7.0.dev6
+Version: 0.7.2
 Summary: Python SDK for the Simudyne Pulse synthetic market data API
 Author-email: Simudyne <support@simudyne.com>
 License-Expression: MIT

{simudyne_pulse-0.7.0.dev6 → simudyne_pulse-0.7.2}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "simudyne-pulse"
-version = "0.7.0-dev.6"
+version = "0.7.2"
 description = "Python SDK for the Simudyne Pulse synthetic market data API"
 readme = "README.md"
 license = "MIT"

{simudyne_pulse-0.7.0.dev6 → simudyne_pulse-0.7.2}/src/simudyne/client.py

@@ -36,8 +36,6 @@ class PulseABM:
         from simudyne.resources.profile import ProfileResource
         from simudyne.resources.api_keys import ApiKeysResource
         from simudyne.resources.data import DataResource
-        from simudyne.resources.fix import FixResource
-        from simudyne.resources.fm import FmResource
         from simudyne.resources.simulation import SimulationResource
         from simudyne.resources.simulator_gym import SimulatorGymResource
         from simudyne.resources.validation import ValidationResource
@@ -45,8 +43,6 @@ class PulseABM:
         self.profile = ProfileResource(self)
         self.api_keys = ApiKeysResource(self)
         self.data = DataResource(self)
-        self.fix = FixResource(self)
-        self.fm = FmResource(self)
         self.simulation = SimulationResource(self)
         self.simulator_gym = SimulatorGymResource(self)
         self.validation = ValidationResource(self)

{simudyne_pulse-0.7.0.dev6 → simudyne_pulse-0.7.2}/src/simudyne/resources/data.py

@@ -11,7 +11,6 @@ class DataResource:
         exchange: str | None = None,
         provider: str | None = None,
         date: str | None = None,
-        q: str | None = None,
         limit: int | None = None,
         offset: int | None = None,
     ):
@@ -32,8 +31,6 @@ class DataResource:
             Calibration date "YYYY-MM-DD". Keeps only instruments
             calibrated on that date, and narrows each instrument's
             available_dates to it.
-        q : str, optional
-            Case-insensitive substring match over ticker and company name.
         limit : int, optional
             Max instruments to return.
         offset : int, optional
@@ -49,7 +46,6 @@ class DataResource:
                 "exchange": exchange,
                 "provider": provider,
                 "date": date,
-                "q": q,
                 "limit": limit,
                 "offset": offset,
             }.items() if v is not None

{simudyne_pulse-0.7.0.dev6 → simudyne_pulse-0.7.2}/src/simudyne/resources/simulation.py

@@ -21,7 +21,6 @@ RESULTS_PATH = "/simulation/results"
 CACHED_PATH = "/simulation/cached"
 SAMPLE_PATH = "/simulation/sample"
 CALIBRATE_PATH = "/calibrate"
-LRM_PATH = "/simulation/lrm/run"
 
 # Available market scenarios
 SCENARIOS = {
@@ -319,17 +318,12 @@ class SimulationResource:
             payload["batch_size"] = batch_size
         return self._pro_request("POST", CALIBRATE_PATH, json=payload)
 
-    def get_jobs(self, limit: int = 100):
-        """Get simulation jobs submitted by the authenticated user, newest first.
+    def get_jobs(self):
+        """Get all simulation jobs submitted by the authenticated user.
 
         Returns a list of jobs with their associated simulation IDs. Use this
         to find job IDs for past runs or to see what simulations are pending.
 
-        Parameters
-        ----------
-        limit : int, default 100
-            Max jobs to return (default 100, max 500)
-
         Returns
         -------
         dict
@@ -339,7 +333,6 @@ class SimulationResource:
             - sim_ids (list): List of simulation IDs in this job
             - created_at (str): Timestamp when job was submitted
             - total (int): Total number of jobs
-            - returned (int): Jobs in this page
 
         Examples
         --------
@@ -350,7 +343,7 @@ class SimulationResource:
         ...     print(f"Job {job['job_id']}: {len(job['sim_ids'])} simulations")
         ...     print(f"  Created: {job['created_at']}")
         """
-        return self._pro_request("GET", JOBS_PATH, params={"limit": limit})
+        return self._pro_request("GET", JOBS_PATH)
 
     def get_job_status(self, job_id: str):
         """Get the status of all simulations in a job.
@@ -479,102 +472,6 @@ class SimulationResource:
         """
         return self._pro_request("GET", f"{JOBS_PATH}/{job_id}/results")
 
-    def get_job_logs(self, job_id: str) -> str:
-        """Fetch the engine run log for one of your jobs, as plain text.
-
-        The worker writes a diagnostic log per job — this is the thing to read
-        when a run fails or finishes with nothing plottable, and the thing to
-        attach when sending a problem to support@simudyne.com.
-
-        Parameters
-        ----------
-        job_id : str
-            The job ID from run() or get_jobs()
-
-        Returns
-        -------
-        str
-            The log text.
-
-        Raises
-        ------
-        PulseAPIError
-            404 when the job does not exist, is not yours, or
-            wrote no log.
-
-        Examples
-        --------
-        >>> status = client.simulation.get_job_status(job_id)
-        >>> if status["has_errors"]:
-        ...     print(client.simulation.get_job_logs(job_id)[:2000])
-        """
-        # Plain text, not JSON — go through the retrying transport directly.
-        url = f"{self._client.base_url}{JOBS_PATH}/{job_id}/logs"
-        response = self._client._request_with_retries("GET", url)
-        return response.text
-
-    def run_lrm(
-        self,
-        symbol: str,
-        cal_date: str,
-        provider: str,
-        exchange: str,
-        order_sizes: list,
-        n_runs: int = 50,
-        seed: int = 42,
-        horizon_mins: int = 60,
-        strategy: str = "vwap",
-        side: str = None,
-    ):
-        """Run a liquidity-risk grid: market impact across a ladder of order sizes.
-
-        One execution algo is built per entry in order_sizes, and all of them
-        share a single baseline, so the cost is ``n_runs * (1 + len(order_sizes))``
-        simulations rather than one baseline per size. Poll the returned job_id
-        through the usual job endpoints.
-
-        Parameters
-        ----------
-        symbol : str
-            Trading symbol (e.g. "700")
-        cal_date : str
-            Calibration date in YYYY-MM-DD format
-        provider : str
-            Data provider (e.g. "omd", "bmll")
-        exchange : str
-            Exchange protocol (e.g. "hkex_securities")
-        order_sizes : list
-            Order sizes in LOTS — one algo per entry
-        n_runs : int, default 50
-            Monte Carlo runs per arm (default 50)
-        seed : int, default 42
-            Random seed (default 42)
-        horizon_mins : int, default 60
-            Execution horizon in minutes (default 60)
-        strategy : str, default 'vwap'
-            "vwap" or "twap" (default "vwap")
-        side : str, optional
-            "buy" or "sell"; defaults to the sign of each order size
-
-        Returns
-        -------
-        dict with job_id and the queued sim_ids
-        """
-        payload = {
-            "symbol": symbol,
-            "cal_date": cal_date,
-            "provider": provider,
-            "exchange": exchange,
-            "order_sizes": order_sizes,
-            "n_runs": n_runs,
-            "seed": seed,
-            "horizon_mins": horizon_mins,
-            "strategy": strategy,
-        }
-        if side is not None:
-            payload["side"] = side
-        return self._pro_request("POST", LRM_PATH, json=payload)
-
     def list_sim_files(self, sim_id: str):
         """List available files for a specific simulation.
 

simudyne_pulse-0.7.2/src/simudyne/resources/validation.py

@@ -0,0 +1,299 @@
+"""
+Validation Resource for the Pulse SDK.
+
+This module provides methods for validating simulation quality by comparing
+simulated LOB data against historical data using distributional metrics,
+impact response analysis, and FID scores.
+
+Workflow:
+    1. Submit a validation job with run() -> returns job_id
+    2. Poll status with get_job(job_id) or use run_pipeline() for blocking
+    3. View results including distances and plots
+    4. List past jobs with list_jobs()
+"""
+
+import time
+import base64
+
+
+RUN_PATH = "/validation/run"
+JOBS_PATH = "/validation/jobs"
+
+
+class ValidationResource:
+    def __init__(self, client):
+        self._client = client
+
+    def run(
+        self,
+        symbol: str,
+        date: str,
+        sim_ids: list[str],
+        ticksize: float = 1.0,
+        run_metrics: bool = True,
+        run_impact: bool = False,
+        run_fid: bool = False,
+        n_levels: int = 10,
+        rescale_volumes: bool = True,
+        lot_size: int = 1,
+        run_stylised_facts: bool | None = None,
+    ) -> dict:
+        """Submit a validation job.
+
+        Compares simulation output against historical market data using
+        distributional distance metrics (L1, Wasserstein), impact response
+        curves, and FID scores.
+
+        Historical data is fetched automatically from GCS based on symbol and date.
+        Simulation data is fetched from each sim_id's sim_data.parquet in GCS.
+
+        Parameters
+        ----------
+        symbol : str
+            Trading symbol (e.g. "700.HK")
+        date : str
+            Calibration date in YYYY-MM-DD format (e.g. "2025-09-01")
+        sim_ids : list[str]
+            List of simulation IDs to validate (max 25)
+        ticksize : float, default 1.0
+            Tick size for the symbol
+        run_metrics : bool, default True
+            Compute L1/Wasserstein distributional distances
+        run_impact : bool, default False
+            Compute impact response curves
+        run_fid : bool, default False
+            Compute Frechet Inception Distance
+        run_stylised_facts : bool, optional
+            Compute stylised facts (autocorrelation, heavy
+            tails, volatility clustering). Left unset it is omitted from the
+            request, so the API applies your tier's default — demo accounts
+            get them, pro accounts do not.
+        n_levels : int, default 10
+            Number of L2 book levels to use
+        rescale_volumes : bool, default True
+            Multiply simulated L2 size columns by lot_size
+        lot_size : int, default 1
+            Lot size multiplier for volume rescaling
+
+        Returns
+        -------
+        dict with job_id, status, message
+        """
+        # run_metrics/run_impact/run_fid keep sending their long-standing values
+        # so existing callers see no change. run_stylised_facts is omitted when
+        # unset, letting the API apply the tier default; sending False would opt
+        # demo accounts out of results that tier is meant to return.
+        config = {
+            "run_metrics": run_metrics,
+            "run_impact": run_impact,
+            "run_fid": run_fid,
+            "n_levels": n_levels,
+            "rescale_volumes": rescale_volumes,
+            "lot_size": lot_size,
+        }
+        if run_stylised_facts is not None:
+            config["run_stylised_facts"] = run_stylised_facts
+
+        payload = {
+            "symbol": symbol,
+            "date": date,
+            "sim_ids": sim_ids,
+            "ticksize": ticksize,
+            "config": config,
+        }
+        return self._client._request("POST", RUN_PATH, json=payload)
+
+    def get_job(self, job_id: str) -> dict:
+        """Get validation job status and results.
+
+        Parameters
+        ----------
+        job_id : str
+            The job ID returned by run()
+
+        Returns
+        -------
+        dict with
+            - status: "pending", "running", "completed", or "failed"
+            - distances: dict of {metric: {l1: [...], w: [...]}} (when completed)
+            - metadata: dict with run parameters, including which passes ran
+            - error: error message (when failed)
+
+            Demo-tier accounts additionally receive the numbers derived from the
+            historical data, which are withheld at the pro tier:
+
+            - distributions: per-metric historical vs simulated histograms
+            - impact_response: impact response curves as numbers, historical and
+            one block per sim run
+            - stylised_facts: historical and one block per sim run
+            - fid_scores: one score per sim run (None where not computable)
+        """
+        return self._client._request("GET", f"{JOBS_PATH}/{job_id}")
+
+    def list_jobs(self, limit: int = 50) -> dict:
+        """List validation jobs for the current user.
+
+        Parameters
+        ----------
+        limit : int, default 50
+            Max number of jobs to return (default 50, max 200)
+
+        Returns
+        -------
+        dict with jobs list and total count
+        """
+        return self._client._request("GET", JOBS_PATH, params={"limit": limit})
+
+    def run_pipeline(
+        self,
+        symbol: str,
+        date: str,
+        sim_ids: list[str],
+        ticksize: float = 1.0,
+        run_metrics: bool = True,
+        run_impact: bool = False,
+        run_fid: bool = False,
+        n_levels: int = 10,
+        rescale_volumes: bool = True,
+        lot_size: int = 1,
+        poll_interval: float = 3.0,
+        timeout: float = 600.0,
+        run_stylised_facts: bool | None = None,
+    ) -> dict:
+        """Submit a validation job and block until it completes.
+
+        Combines run() + polling get_job() into a single call.
+        Prints progress to stderr.
+
+        Parameters
+        ----------
+        symbol : str
+            Trading symbol (e.g. "700.HK")
+        date : str
+            Calibration date in YYYY-MM-DD format
+        sim_ids : list[str]
+            List of simulation IDs to validate (max 25)
+        ticksize : float, default 1.0
+            Tick size for the symbol
+        run_metrics : bool, default True
+            Compute L1/Wasserstein distributional distances
+        run_impact : bool, default False
+            Compute impact response curves
+        run_fid : bool, default False
+            Compute Frechet Inception Distance
+        run_stylised_facts : bool, optional
+            Compute stylised facts (autocorrelation, heavy
+            tails, volatility clustering). Left unset it is omitted from the
+            request, so the API applies your tier's default — demo accounts
+            get them, pro accounts do not.
+        n_levels : int, default 10
+            Number of L2 book levels to use
+        rescale_volumes : bool, default True
+            Multiply simulated L2 size columns by lot_size
+        lot_size : int, default 1
+            Lot size multiplier for volume rescaling
+        poll_interval : float, default 3.0
+            Seconds between status checks (default 3)
+        timeout : float, default 600.0
+            Max seconds to wait (default 600)
+
+        Returns
+        -------
+        dict with full validation results (distances, plots, metadata)
+
+        Raises
+        ------
+        RuntimeError
+            If the validation job fails
+        TimeoutError
+            If the job doesn't complete within timeout
+        """
+        import sys
+
+        job = self.run(
+            symbol=symbol,
+            date=date,
+            sim_ids=sim_ids,
+            ticksize=ticksize,
+            run_metrics=run_metrics,
+            run_impact=run_impact,
+            run_fid=run_fid,
+            run_stylised_facts=run_stylised_facts,
+            n_levels=n_levels,
+            rescale_volumes=rescale_volumes,
+            lot_size=lot_size,
+        )
+        job_id = job["job_id"]
+        print(f"Validation job submitted: {job_id}", file=sys.stderr)
+
+        start = time.time()
+        while True:
+            result = self.get_job(job_id)
+            status = result["status"]
+
+            if status == "completed":
+                elapsed = time.time() - start
+                print(f"Completed in {elapsed:.1f}s", file=sys.stderr)
+                return result
+            elif status == "failed":
+                raise RuntimeError(f"Validation failed: {result.get('error')}")
+
+            if time.time() - start > timeout:
+                raise TimeoutError(f"Validation job {job_id} timed out after {timeout}s")
+
+            time.sleep(poll_interval)
+
+    def display_plots(self, result: dict) -> "PlotDisplay":
+        """Return a PlotDisplay object for displaying validation plots.
+
+        Usage:
+            plots = client.validation.display_plots(result)
+            plots.distributions()    # show distribution histograms
+            plots.distances()        # show spider plots
+            plots.impact_response()  # show impact response plots
+
+        Parameters
+        ----------
+        result : dict
+            The result dict from run_pipeline() or get_job()
+        """
+        return PlotDisplay(result)
+
+
+class PlotDisplay:
+    """Displays categorized validation plots inline in Jupyter notebooks."""
+
+    def __init__(self, result: dict):
+        plots = result.get("plots") or {}
+        self._distributions = plots.get("distributions", [])
+        self._distances = plots.get("distances", [])
+        self._impact_response = plots.get("impact_response", [])
+
+    def _show(self, plot_list, title):
+        from IPython.display import display, Image
+
+        if not plot_list:
+            print(f"No {title} plots available")
+            return
+
+        for plot in plot_list:
+            print(f"\n--- {plot['name']} ---")
+            display(Image(data=base64.b64decode(plot["content_base64"])))
+
+    def distributions(self):
+        """Display distribution histogram plots."""
+        self._show(self._distributions, "distribution")
+
+    def distances(self):
+        """Display spider plots (L1 and Wasserstein distances)."""
+        self._show(self._distances, "distance")
+
+    def impact_response(self):
+        """Display impact response plots."""
+        self._show(self._impact_response, "impact response")
+
+    def all(self):
+        """Display all plots."""
+        self.distances()
+        self.distributions()
+        self.impact_response()

{simudyne_pulse-0.7.0.dev6 → simudyne_pulse-0.7.2/src/simudyne_pulse.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: simudyne-pulse
-Version: 0.7.0.dev6
+Version: 0.7.2
 Summary: Python SDK for the Simudyne Pulse synthetic market data API
 Author-email: Simudyne <support@simudyne.com>
 License-Expression: MIT

{simudyne_pulse-0.7.0.dev6 → simudyne_pulse-0.7.2}/src/simudyne_pulse.egg-info/SOURCES.txt

@@ -7,8 +7,6 @@ src/simudyne/exceptions.py
 src/simudyne/resources/__init__.py
 src/simudyne/resources/api_keys.py
 src/simudyne/resources/data.py
-src/simudyne/resources/fix.py
-src/simudyne/resources/fm.py
 src/simudyne/resources/historical.py
 src/simudyne/resources/profile.py
 src/simudyne/resources/simulation.py
@@ -19,6 +17,4 @@ src/simudyne_pulse.egg-info/SOURCES.txt
 src/simudyne_pulse.egg-info/dependency_links.txt
 src/simudyne_pulse.egg-info/requires.txt
 src/simudyne_pulse.egg-info/top_level.txt
-tests/test_new_resources.py
-tests/test_simulator_gym.py
-tests/test_validation.py
+tests/test_simulator_gym.py

simudyne_pulse-0.7.0.dev6/src/simudyne/resources/fix.py

@@ -1,32 +0,0 @@
-"""
-FIX Resource for the Pulse SDK.
-
-Pulse simulations can be consumed over the FIX protocol by a market-data
-session instead of the REST download endpoints. The FIX connection itself is
-provisioned per organisation (see the FIX docs page); this resource covers
-what the SDK can usefully do about it — reporting what your account has run
-over FIX.
-"""
-
-USAGE_PATH = "/fix/usage"
-
-
-class FixResource:
-    def __init__(self, client):
-        self._client = client
-
-    def usage(self) -> dict:
-        """FIX simulation statistics: what was run over FIX, and how often.
-
-        Returns
-        -------
-        dict with ``total_runs``, ``simulated_seconds``, and ``runs`` — one
-            entry per distinct configuration, each with its ``symbol``,
-            ``cal_date``, ``scenario`` and ``runs`` count.
-
-        Examples
-        --------
-        >>> usage = client.fix.usage()
-        >>> print(f"{usage['total_runs']} FIX runs")
-        """
-        return self._client._request("GET", USAGE_PATH)