electrochem: Fix mpt float parsing using locales. (#99)

* Fix mpt parsing using locales.

* Fix locale ordering.

* Try fix locale-gen.

* black

.github/workflows/test-job/action.yml

@@ -3,7 +3,12 @@ description: "Test the package with pytest."
 runs:
   using: "composite"
   steps:
-    - run: yadg --version
+    - name: Prepare locales
+      run: sudo locale-gen de_DE.UTF-8 || echo windows
       shell: bash
-    - run: pytest -vv
+    - name: Print yadg version
+      run: yadg --version
+      shell: bash
+    - name: Run pytest
+      run: pytest -vv
       shell: bash

src/yadg/parsers/electrochem/eclabmpt.py

@@ -33,6 +33,7 @@
 """
 import re
 import logging
+import locale
 from collections import defaultdict
 from ...dgutils.dateutils import str_to_uts
 from .eclabtechniques import get_resolution, technique_params, param_from_key
@@ -154,14 +155,29 @@ def _process_header(lines: list[str], timezone: str) -> tuple[dict, list, dict]:
     settings_lines = sections[2].split("\n")
     technique, params_keys = technique_params(technique, settings_lines)
     params = settings_lines[-len(params_keys) :]
+
+    # Get the locale
+    old_loc = locale.getlocale(category=locale.LC_NUMERIC)
+    ewe_ctrl_re = re.compile(r"Ewe ctrl range : min = (?P<min>.+), max = (?P<max>.+)")
+    ewe_ctrl_match = ewe_ctrl_re.search("\n".join(settings_lines))
+    for loc in [old_loc, "de_DE.UTF-8", "en_GB.UTF-8", "en_US.UTF-8"]:
+        try:
+            locale.setlocale(locale.LC_NUMERIC, locale=loc)
+            locale.atof(ewe_ctrl_match["min"].split()[0])
+            locale.atof(ewe_ctrl_match["max"].split()[0])
+            logging.debug(f"The locale of the current file is '{loc}'.")
+            break
+        except ValueError:
+            logging.debug(f"Could not parse Ewe ctrl using locale '{loc}'.")
+
     # The sequence param columns are always allocated 20 characters.
     n_sequences = int(len(params[0]) / 20)
     params_values = []
     for seq in range(1, n_sequences):
         values = []
         for param in params:
             try:
-                val = float(param[seq * 20 : (seq + 1) * 20])
+                val = locale.atof(param[seq * 20 : (seq + 1) * 20])
             except ValueError:
                 val = param[seq * 20 : (seq + 1) * 20].strip()
             values.append(val)
@@ -192,6 +208,7 @@ def _process_header(lines: list[str], timezone: str) -> tuple[dict, list, dict]:
         loops = {"n_loops": n_loops, "indexes": indexes}
     settings = {
         "posix_timestamp": uts,
+        "locale": loc,
         "technique": technique,
         "raw": "\n".join(lines),
     }
@@ -228,7 +245,7 @@ def _process_data(lines: list[str], Eranges: list[float], Iranges: list[float])
         datapoint = {}
         for col, val, unit in list(zip(columns, values, units)):
             if unit is None:
-                ival = int(float(val))
+                ival = int(locale.atof(val))
                 if col == "I Range":
                     datapoint[col] = param_from_key("I_range", ival)
                     Irange = param_from_key("I_range", ival, to_str=False)
@@ -252,7 +269,7 @@ def _process_data(lines: list[str], Eranges: list[float], Iranges: list[float])
             Irange = param_from_key("I_range", Irstr, to_str=False)
         for col, val, unit in list(zip(columns, values, units)):
             try:
-                val = float(val)
+                val = locale.atof(val)
             except ValueError:
                 val = val.strip()
             if unit is None:
@@ -299,6 +316,8 @@ def process(
     header_lines = lines[: nb_header_lines - 3]
     data_lines = lines[nb_header_lines - 3 :]
     settings, params, loops = {}, [], {}
+    # Store current LC_NUMERIC before we do anything:
+    old_loc = locale.getlocale(category=locale.LC_NUMERIC)
     if nb_header_lines <= 3:
         logger.warning("Header contains no settings and hence no timestamp.")
         start_time = 0.0
@@ -352,4 +371,6 @@ def process(
         uts = start_time + d["time"]["n"]
         d["technique"] = settings["technique"]
         timesteps.append({"fn": fn, "uts": uts, "raw": d})
+    # reset to original LC_NUMERIC
+    locale.setlocale(category=locale.LC_NUMERIC, locale=old_loc)
     return timesteps, metadata, fulldate

tests/test_electrochem.py

@@ -387,6 +387,23 @@
                 },
             },
         ),
+        (  # ts21 - mb_67.mpt
+            {
+                "case": "mb_67.mpt",
+                "encoding": "windows-1252",
+                "parameters": {"filetype": "eclab.mpt"},
+            },
+            {
+                "nsteps": 1,
+                "step": 0,
+                "nrows": 33,
+                "point": 0,
+                "pars": {
+                    "uts": {"value": 1670510213.355},
+                    "Ewe": {"value": 2.3278546, "unit": "V", "sigma": 7.5e-05},
+                },
+            },
+        ),
     ],
 )
 def test_datagram_from_eclab(input, ts, datadir):