From 20176f4b981a904052e0fe8180a0a6f8deaec473 Mon Sep 17 00:00:00 2001
From: "Edgar P. Burkhart" <git@edgarpierre.fr>
Date: Thu, 31 Mar 2022 10:24:38 +0200
Subject: [PATCH] Time-series pre-processing

---
 data/config.ini       |  1 +
 data/data/.gitignore  |  1 +
 data/processing/ts.py | 42 ++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 44 insertions(+)
 create mode 100644 data/processing/ts.py

diff --git a/data/config.ini b/data/config.ini
index 3b784bc..2ca1c99 100644
--- a/data/config.ini
+++ b/data/config.ini
@@ -5,6 +5,7 @@ hires=bathyhires.dat
 hstru=Hstru.dat
 poro=Poro.dat
 psize=Psize.dat
+raw_ts=201702281700.raw
 hires_step=0.5
 
 [out]
diff --git a/data/data/.gitignore b/data/data/.gitignore
index ef83881..47293d7 100644
--- a/data/data/.gitignore
+++ b/data/data/.gitignore
@@ -1 +1,2 @@
 *.xyz
+*.raw
diff --git a/data/processing/ts.py b/data/processing/ts.py
new file mode 100644
index 0000000..428da00
--- /dev/null
+++ b/data/processing/ts.py
@@ -0,0 +1,42 @@
+import argparse
+import configparser
+import logging
+import pathlib
+
+import numpy as np
+
+parser = argparse.ArgumentParser(description="Pre-process time-series")
+parser.add_argument("-v", "--verbose", action="count", default=0)
+parser.add_argument("-c", "--config", default="config.ini")
+args = parser.parse_args()
+
+logging.basicConfig(level=max((10, 20 - 10 * args.verbose)))
+log = logging.getLogger("bathy")
+
+log.info("Starting time-series pre-processing")
+config = configparser.ConfigParser()
+config.read(args.config)
+
+inp_root = pathlib.Path(config.get("inp", "root"))
+out_root = pathlib.Path(config.get("out", "root"))
+
+inp_ts = inp_root.joinpath(config.get("inp", "raw_ts"))
+out_ts = out_root.joinpath("ts.dat")
+
+raw_ts = np.loadtxt(
+    inp_ts,
+    dtype=[("state", int), ("z", float), ("y", float), ("x", float)],
+    delimiter=",",
+    max_rows=2304,
+)
+log.debug(f"{raw_ts=}")
+
+if (errs := np.count_nonzero(raw_ts["state"])) != 0:
+    log.warning(f"{errs} transmission errors!")
+    log.debug(f"{dict(zip(*np.unique(raw_ts['state'], return_counts=True)))}")
+
+t = np.linspace(0, 30 * 60, 2305)[:-1]
+log.debug(f"{t=}")
+
+log.info(f"Saving timeseries to '{out_ts}'")
+np.savetxt(out_ts, np.stack((t, raw_ts["z"]), axis=1))