From aba5750d55c0ac495ab0e7e1b3af8ca33850346f Mon Sep 17 00:00:00 2001
From: GeorgWa <wallmann@biochem.mpg.de>
Date: Thu, 23 Nov 2023 00:45:41 +0100
Subject: [PATCH 1/3] FEAT using mp for file loading under linux

---
 alpharaw/thermo.py | 20 +++++++++-----------
 1 file changed, 9 insertions(+), 11 deletions(-)

diff --git a/alpharaw/thermo.py b/alpharaw/thermo.py
index d0c98f0..a09fc75 100644
--- a/alpharaw/thermo.py
+++ b/alpharaw/thermo.py
@@ -118,7 +118,7 @@ class ThermoRawData(MSData_Base):
     def __init__(self, 
             centroided : bool = True,
             process_count : int = 10,
-            mp_batch_size : int = 10000,
+            mp_batch_size : int = 5000,
             **kwargs):
         """
         Parameters
@@ -148,17 +148,15 @@ def _import(self,
         first_spectrum_number = rawfile.FirstSpectrumNumber
         last_spectrum_number = rawfile.LastSpectrumNumber
 
-        if platform.system() != 'Linux':
-            batches = np.arange(first_spectrum_number, last_spectrum_number+1, self.mp_batch_size)
-            batches = np.append(batches, last_spectrum_number+1)
+        mode = 'spawn' if platform.system() != 'Linux' else 'forkserver'
+            
+        batches = np.arange(first_spectrum_number, last_spectrum_number+1, self.mp_batch_size)
+        batches = np.append(batches, last_spectrum_number+1)
 
-            # use multiprocessing to load batches
-            _import_batch_partial = partial(_import_batch, raw_file_path, self.centroided)
-            with mp.get_context("spawn").Pool(processes = self.process_count) as pool:
-                batches = list(tqdm(pool.imap(_import_batch_partial, zip(batches[:-1], batches[1:]))))
-
-        else:
-            batches = [_import_batch(raw_file_path, self.centroided, (first_spectrum_number, last_spectrum_number+1))]
+        # use multiprocessing to load batches
+        _import_batch_partial = partial(_import_batch, raw_file_path, self.centroided)
+        with mp.get_context(mode).Pool(processes = self.process_count) as pool:
+            batches = list(tqdm(pool.imap(_import_batch_partial, zip(batches[:-1], batches[1:]))))
 
         # collect peak indices
         _peak_indices = np.concatenate([batch['_peak_indices'] for batch in batches])        

From 8e39a1baefa4b55d8436e3e897caccf307998d43 Mon Sep 17 00:00:00 2001
From: GeorgWa <wallmann@biochem.mpg.de>
Date: Thu, 23 Nov 2023 01:00:34 +0100
Subject: [PATCH 2/3] FIX disable warnings

---
 alpharaw/__init__.py | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/alpharaw/__init__.py b/alpharaw/__init__.py
index 29b5952..8b6b502 100644
--- a/alpharaw/__init__.py
+++ b/alpharaw/__init__.py
@@ -1,5 +1,8 @@
 #!python
 
+import warnings
+warnings.filterwarnings("ignore")
+
 def register_readers():
     from .ms_data_base import ms_reader_provider
     from .legacy_msdata import mgf

From 930c3fa9f30d62ec32f6f58ca2e05ae1826f2986 Mon Sep 17 00:00:00 2001
From: GeorgWa <wallmann@biochem.mpg.de>
Date: Thu, 23 Nov 2023 01:20:18 +0100
Subject: [PATCH 3/3] FIX copy memory explicitly

---
 alpharaw/thermo.py | 23 +++++++++++++----------
 1 file changed, 13 insertions(+), 10 deletions(-)

diff --git a/alpharaw/thermo.py b/alpharaw/thermo.py
index a09fc75..edb0861 100644
--- a/alpharaw/thermo.py
+++ b/alpharaw/thermo.py
@@ -98,18 +98,21 @@ def _import_batch(
                 isolation_mz_uppers.append(isolation_center + isolation_width / 2)
                 precursor_charges.append(charge)
     rawfile.Close()
+
+    # copys of numpy arrays are needed to move them explicitly to cpython heap
+    # otherwise mono might interfere later
     return {
         '_peak_indices': _peak_indices,
-        'peak_mz': np.concatenate(mz_values),
-        'peak_intensity': np.concatenate(intensity_values),
-        'rt': np.array(rt_values),
-        'precursor_mz': np.array(precursor_mz_values),
-        'precursor_charge': np.array(precursor_charges, dtype=np.int8),
-        'isolation_lower_mz': np.array(isolation_mz_lowers),
-        'isolation_upper_mz': np.array(isolation_mz_uppers),
-        'ms_level': np.array(ms_order_list, dtype=np.int8),
-        'nce': np.array(ce_list, dtype=np.float32),
-        'injection_time': np.array(injection_time_list, dtype=np.float32)
+        'peak_mz': np.concatenate(mz_values).copy(),
+        'peak_intensity': np.concatenate(intensity_values).copy(),
+        'rt': np.array(rt_values).copy(),
+        'precursor_mz': np.array(precursor_mz_values).copy(),
+        'precursor_charge': np.array(precursor_charges, dtype=np.int8).copy(),
+        'isolation_lower_mz': np.array(isolation_mz_lowers).copy(),
+        'isolation_upper_mz': np.array(isolation_mz_uppers).copy(),
+        'ms_level': np.array(ms_order_list, dtype=np.int8).copy(),
+        'nce': np.array(ce_list, dtype=np.float32).copy(),
+        'injection_time': np.array(injection_time_list, dtype=np.float32).copy()
     }
 class ThermoRawData(MSData_Base):
     """