osekit 1.3.0__tar.gz → 1.5.1__tar.gz

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (201) hide show
  1. {osekit-1.3.0 → osekit-1.5.1}/CITATION.cff +1 -1
  2. {osekit-1.3.0 → osekit-1.5.1}/PKG-INFO +1 -1
  3. {osekit-1.3.0 → osekit-1.5.1}/docs/source/coreapi_usage.rst +7 -9
  4. {osekit-1.3.0 → osekit-1.5.1}/docs/source/example_ltas_core.ipynb +1 -1
  5. {osekit-1.3.0 → osekit-1.5.1}/docs/source/example_ltas_public.ipynb +3 -3
  6. {osekit-1.3.0 → osekit-1.5.1}/docs/source/example_multichannel.ipynb +1 -1
  7. {osekit-1.3.0 → osekit-1.5.1}/docs/source/example_multiple_spectrograms_public.ipynb +1 -1
  8. {osekit-1.3.0 → osekit-1.5.1}/docs/source/example_spectrogram.ipynb +4 -4
  9. {osekit-1.3.0 → osekit-1.5.1}/docs/source/examples.rst +2 -2
  10. osekit-1.5.1/docs/source/job.rst +27 -0
  11. {osekit-1.3.0 → osekit-1.5.1}/docs/source/jobs.rst +41 -23
  12. {osekit-1.3.0 → osekit-1.5.1}/docs/source/multichannel.rst +2 -2
  13. {osekit-1.3.0 → osekit-1.5.1}/docs/source/publicapi_usage.rst +7 -7
  14. {osekit-1.3.0 → osekit-1.5.1}/pyproject.toml +1 -1
  15. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/core/audio_data.py +33 -17
  16. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/core/audio_item.py +24 -3
  17. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/core/base_dataset.py +13 -7
  18. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/core/detection.py +25 -2
  19. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/core/event.py +14 -19
  20. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/core/frequency_scale.py +4 -4
  21. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/core/ltas_data.py +4 -4
  22. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/core/spectro_data.py +34 -26
  23. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/core/spectro_dataset.py +6 -6
  24. osekit-1.5.1/src/osekit/job/builder.py +153 -0
  25. osekit-1.5.1/src/osekit/job/config.py +32 -0
  26. osekit-1.5.1/src/osekit/job/job.py +237 -0
  27. osekit-1.5.1/src/osekit/job/scheduler/pbs.py +209 -0
  28. osekit-1.5.1/src/osekit/job/scheduler/scheduler.py +301 -0
  29. osekit-1.5.1/src/osekit/job/scheduler/slurm.py +267 -0
  30. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/public/export_transform.py +7 -7
  31. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/public/project.py +2 -2
  32. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/public/transform.py +8 -8
  33. osekit-1.3.0/tests/sample_dataset/sample1_180701_204453.wav → osekit-1.5.1/tests/__init__.py +0 -0
  34. {osekit-1.3.0 → osekit-1.5.1}/tests/_static/aplose_result.csv +2 -1
  35. osekit-1.5.1/tests/_static/job_status_request_results/pbs.txt +39 -0
  36. osekit-1.5.1/tests/_static/job_status_request_results/slurm_sacct.txt +1 -0
  37. osekit-1.5.1/tests/_static/job_status_request_results/slurm_squeue.txt +1 -0
  38. osekit-1.5.1/tests/_static/retro_compatibility_result.csv +2 -0
  39. {osekit-1.3.0 → osekit-1.5.1}/tests/conftest.py +2 -2
  40. osekit-1.3.0/tests/sample_dataset/sample1_190222_010000.wav → osekit-1.5.1/tests/helpers/__init__.py +0 -0
  41. {osekit-1.3.0 → osekit-1.5.1}/tests/helpers/audio.py +13 -2
  42. osekit-1.5.1/tests/sample_dataset/sample1_190222_010000.wav +0 -0
  43. osekit-1.5.1/tests/sample_dataset/sample1_220998_185532.wav +0 -0
  44. {osekit-1.3.0 → osekit-1.5.1}/tests/test_audio.py +110 -1
  45. {osekit-1.3.0 → osekit-1.5.1}/tests/test_core_api_base.py +45 -25
  46. {osekit-1.3.0 → osekit-1.5.1}/tests/test_detection.py +33 -4
  47. {osekit-1.3.0 → osekit-1.5.1}/tests/test_event.py +31 -0
  48. {osekit-1.3.0 → osekit-1.5.1}/tests/test_export_transform.py +2 -2
  49. osekit-1.5.1/tests/test_job.py +973 -0
  50. {osekit-1.3.0 → osekit-1.5.1}/tests/test_public_api.py +9 -9
  51. {osekit-1.3.0 → osekit-1.5.1}/tests/test_serialization.py +9 -6
  52. {osekit-1.3.0 → osekit-1.5.1}/tests/test_spectro.py +37 -46
  53. {osekit-1.3.0 → osekit-1.5.1}/tests/test_welch.py +25 -0
  54. {osekit-1.3.0 → osekit-1.5.1}/uv.lock +248 -203
  55. osekit-1.3.0/docs/source/job.rst +0 -7
  56. osekit-1.3.0/src/osekit/core_api/__init__.py +0 -12
  57. osekit-1.3.0/src/osekit/public_api/__init__.py +0 -12
  58. osekit-1.3.0/src/osekit/utils/job.py +0 -635
  59. osekit-1.3.0/tests/test_job.py +0 -799
  60. {osekit-1.3.0 → osekit-1.5.1}/.github/dependabot.yml +0 -0
  61. {osekit-1.3.0 → osekit-1.5.1}/.github/workflows/dependabot-auto-merge.yml +0 -0
  62. {osekit-1.3.0 → osekit-1.5.1}/.github/workflows/documentation.yml +0 -0
  63. {osekit-1.3.0 → osekit-1.5.1}/.github/workflows/github_ci.yml +0 -0
  64. {osekit-1.3.0 → osekit-1.5.1}/.github/workflows/publish.yml +0 -0
  65. {osekit-1.3.0 → osekit-1.5.1}/.gitignore +0 -0
  66. {osekit-1.3.0 → osekit-1.5.1}/.pre-commit-config.yaml +0 -0
  67. {osekit-1.3.0 → osekit-1.5.1}/CHANGELOG.md +0 -0
  68. {osekit-1.3.0 → osekit-1.5.1}/CONTRIBUTING.md +0 -0
  69. {osekit-1.3.0 → osekit-1.5.1}/LICENSE +0 -0
  70. {osekit-1.3.0 → osekit-1.5.1}/README.md +0 -0
  71. {osekit-1.3.0 → osekit-1.5.1}/docs/Makefile +0 -0
  72. {osekit-1.3.0 → osekit-1.5.1}/docs/logo/osekit.png +0 -0
  73. {osekit-1.3.0 → osekit-1.5.1}/docs/logo/osekit_mid.png +0 -0
  74. {osekit-1.3.0 → osekit-1.5.1}/docs/logo/osekit_small.png +0 -0
  75. {osekit-1.3.0 → osekit-1.5.1}/docs/logo/osmose.png +0 -0
  76. {osekit-1.3.0 → osekit-1.5.1}/docs/logo/osmose_texte_sombre.png +0 -0
  77. {osekit-1.3.0 → osekit-1.5.1}/docs/logo/osmose_texte_sombre_small.png +0 -0
  78. {osekit-1.3.0 → osekit-1.5.1}/docs/make.bat +0 -0
  79. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/audio_data/ex2.svg +0 -0
  80. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/audio_data/ex_1.svg +0 -0
  81. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/audio_data/ex_2.svg +0 -0
  82. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/custom.css +0 -0
  83. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/data_structure/data_structure.svg +0 -0
  84. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/data_structure/optimized_IO.svg +0 -0
  85. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/detections/aplose_results.csv +0 -0
  86. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/detections/detection_labelled.png +0 -0
  87. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/detections/detection_unlabelled.png +0 -0
  88. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/detections/label_anchors.png +0 -0
  89. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/favicon.ico +0 -0
  90. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/ltas/ltas.svg +0 -0
  91. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/sample_audio/id/1829.wav +0 -0
  92. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/sample_audio/id/1830.wav +0 -0
  93. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/sample_audio/id/1832.wav +0 -0
  94. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/sample_audio/id/1833.wav +0 -0
  95. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/sample_audio/id/1834.wav +0 -0
  96. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/sample_audio/id/1835.wav +0 -0
  97. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/sample_audio/id/1836.wav +0 -0
  98. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/sample_audio/id/1837.wav +0 -0
  99. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/sample_audio/id/1838.wav +0 -0
  100. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/sample_audio/multichannel/multichannel_220925_223450.wav +0 -0
  101. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/sample_audio/timestamped/sample_220925_223450.wav +0 -0
  102. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/sample_audio/timestamped/sample_220925_223500.wav +0 -0
  103. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/sample_audio/timestamped/sample_220925_223510.wav +0 -0
  104. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/sample_audio/timestamped/sample_220925_223520.wav +0 -0
  105. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/sample_audio/timestamped/sample_220925_223530.wav +0 -0
  106. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/sample_audio/timestamped/sample_220925_223600.wav +0 -0
  107. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/sample_audio/timestamped/sample_220925_223610.wav +0 -0
  108. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/sample_audio/timestamped/sample_220925_223620.wav +0 -0
  109. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/sample_audio/timestamped/sample_220925_223630.wav +0 -0
  110. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/sample_audio/timestamped/sample_220925_223640.wav +0 -0
  111. {osekit-1.3.0 → osekit-1.5.1}/docs/source/_static/spectrograms/frequency_scale.svg +0 -0
  112. {osekit-1.3.0 → osekit-1.5.1}/docs/source/api.rst +0 -0
  113. {osekit-1.3.0 → osekit-1.5.1}/docs/source/aplose.rst +0 -0
  114. {osekit-1.3.0 → osekit-1.5.1}/docs/source/audiodata.rst +0 -0
  115. {osekit-1.3.0 → osekit-1.5.1}/docs/source/audiodataset.rst +0 -0
  116. {osekit-1.3.0 → osekit-1.5.1}/docs/source/audiofile.rst +0 -0
  117. {osekit-1.3.0 → osekit-1.5.1}/docs/source/audiofilemanager.rst +0 -0
  118. {osekit-1.3.0 → osekit-1.5.1}/docs/source/audioitem.rst +0 -0
  119. {osekit-1.3.0 → osekit-1.5.1}/docs/source/basedata.rst +0 -0
  120. {osekit-1.3.0 → osekit-1.5.1}/docs/source/basedataset.rst +0 -0
  121. {osekit-1.3.0 → osekit-1.5.1}/docs/source/basefile.rst +0 -0
  122. {osekit-1.3.0 → osekit-1.5.1}/docs/source/baseitem.rst +0 -0
  123. {osekit-1.3.0 → osekit-1.5.1}/docs/source/conf.py +0 -0
  124. {osekit-1.3.0 → osekit-1.5.1}/docs/source/contributing.rst +0 -0
  125. {osekit-1.3.0 → osekit-1.5.1}/docs/source/coreapi.rst +0 -0
  126. {osekit-1.3.0 → osekit-1.5.1}/docs/source/coreapi_home.rst +0 -0
  127. {osekit-1.3.0 → osekit-1.5.1}/docs/source/coreapi_introduction.rst +0 -0
  128. {osekit-1.3.0 → osekit-1.5.1}/docs/source/detection.rst +0 -0
  129. {osekit-1.3.0 → osekit-1.5.1}/docs/source/event.rst +0 -0
  130. {osekit-1.3.0 → osekit-1.5.1}/docs/source/example_aplose_result.ipynb +0 -0
  131. {osekit-1.3.0 → osekit-1.5.1}/docs/source/example_ltas.rst +0 -0
  132. {osekit-1.3.0 → osekit-1.5.1}/docs/source/example_multiple_spectrograms.rst +0 -0
  133. {osekit-1.3.0 → osekit-1.5.1}/docs/source/example_multiple_spectrograms_core.ipynb +0 -0
  134. {osekit-1.3.0 → osekit-1.5.1}/docs/source/example_multiple_spectrograms_id.rst +0 -0
  135. {osekit-1.3.0 → osekit-1.5.1}/docs/source/example_multiple_spectrograms_id_core.ipynb +0 -0
  136. {osekit-1.3.0 → osekit-1.5.1}/docs/source/example_multiple_spectrograms_id_public.ipynb +0 -0
  137. {osekit-1.3.0 → osekit-1.5.1}/docs/source/example_reshaping_multiple_files.rst +0 -0
  138. {osekit-1.3.0 → osekit-1.5.1}/docs/source/example_reshaping_multiple_files_core.ipynb +0 -0
  139. {osekit-1.3.0 → osekit-1.5.1}/docs/source/example_reshaping_multiple_files_public.ipynb +0 -0
  140. {osekit-1.3.0 → osekit-1.5.1}/docs/source/example_reshaping_one_file.ipynb +0 -0
  141. {osekit-1.3.0 → osekit-1.5.1}/docs/source/export_transform.rst +0 -0
  142. {osekit-1.3.0 → osekit-1.5.1}/docs/source/frequencyscale.rst +0 -0
  143. {osekit-1.3.0 → osekit-1.5.1}/docs/source/index.rst +0 -0
  144. {osekit-1.3.0 → osekit-1.5.1}/docs/source/installation.rst +0 -0
  145. {osekit-1.3.0 → osekit-1.5.1}/docs/source/instrument.rst +0 -0
  146. {osekit-1.3.0 → osekit-1.5.1}/docs/source/ltasdata.rst +0 -0
  147. {osekit-1.3.0 → osekit-1.5.1}/docs/source/multiprocessing.rst +0 -0
  148. {osekit-1.3.0 → osekit-1.5.1}/docs/source/normalization.rst +0 -0
  149. {osekit-1.3.0 → osekit-1.5.1}/docs/source/project.rst +0 -0
  150. {osekit-1.3.0 → osekit-1.5.1}/docs/source/publicapi.rst +0 -0
  151. {osekit-1.3.0 → osekit-1.5.1}/docs/source/spectrodata.rst +0 -0
  152. {osekit-1.3.0 → osekit-1.5.1}/docs/source/spectrodataset.rst +0 -0
  153. {osekit-1.3.0 → osekit-1.5.1}/docs/source/spectrofile.rst +0 -0
  154. {osekit-1.3.0 → osekit-1.5.1}/docs/source/spectroitem.rst +0 -0
  155. {osekit-1.3.0 → osekit-1.5.1}/docs/source/transform.rst +0 -0
  156. {osekit-1.3.0 → osekit-1.5.1}/docs/source/usage.rst +0 -0
  157. {osekit-1.3.0 → osekit-1.5.1}/docs/source/utils.rst +0 -0
  158. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/__init__.py +0 -0
  159. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/audio_backend/audio_backend.py +0 -0
  160. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/audio_backend/audio_file_manager.py +0 -0
  161. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/audio_backend/mseed_backend.py +0 -0
  162. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/audio_backend/soundfile_backend.py +0 -0
  163. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/config.py +0 -0
  164. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/core/__init__.py +0 -0
  165. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/core/audio_dataset.py +0 -0
  166. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/core/audio_file.py +0 -0
  167. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/core/base_data.py +0 -0
  168. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/core/base_file.py +0 -0
  169. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/core/base_item.py +0 -0
  170. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/core/instrument.py +0 -0
  171. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/core/json_serializer.py +0 -0
  172. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/core/ltas_dataset.py +0 -0
  173. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/core/spectro_file.py +0 -0
  174. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/core/spectro_item.py +0 -0
  175. {osekit-1.3.0/src/osekit/public → osekit-1.5.1/src/osekit/job}/__init__.py +0 -0
  176. {osekit-1.3.0/src/osekit/utils → osekit-1.5.1/src/osekit/job/scheduler}/__init__.py +0 -0
  177. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/logging_config.yaml +0 -0
  178. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/logging_context.py +0 -0
  179. {osekit-1.3.0/tests → osekit-1.5.1/src/osekit/public}/__init__.py +0 -0
  180. {osekit-1.3.0/tests/helpers → osekit-1.5.1/src/osekit/utils}/__init__.py +0 -0
  181. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/utils/aplose.py +0 -0
  182. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/utils/audio.py +0 -0
  183. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/utils/core.py +0 -0
  184. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/utils/deserialization.py +0 -0
  185. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/utils/multiprocess.py +0 -0
  186. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/utils/path.py +0 -0
  187. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/utils/plot.py +0 -0
  188. {osekit-1.3.0 → osekit-1.5.1}/src/osekit/utils/timestamp.py +0 -0
  189. {osekit-1.3.0 → osekit-1.5.1}/tests/_static/minimal_detector_result.csv +0 -0
  190. {osekit-1.3.0 → osekit-1.5.1}/tests/helpers/dummy.py +0 -0
  191. /osekit-1.3.0/tests/sample_dataset/sample1_220998_185532.wav → /osekit-1.5.1/tests/sample_dataset/sample1_180701_204453.wav +0 -0
  192. {osekit-1.3.0 → osekit-1.5.1}/tests/test_audio_file_manager.py +0 -0
  193. {osekit-1.3.0 → osekit-1.5.1}/tests/test_files.py +0 -0
  194. {osekit-1.3.0 → osekit-1.5.1}/tests/test_frequency_scales.py +0 -0
  195. {osekit-1.3.0 → osekit-1.5.1}/tests/test_instrument.py +0 -0
  196. {osekit-1.3.0 → osekit-1.5.1}/tests/test_logging.py +0 -0
  197. {osekit-1.3.0 → osekit-1.5.1}/tests/test_mseed.py +0 -0
  198. {osekit-1.3.0 → osekit-1.5.1}/tests/test_multiprocessing.py +0 -0
  199. {osekit-1.3.0 → osekit-1.5.1}/tests/test_permissions.py +0 -0
  200. {osekit-1.3.0 → osekit-1.5.1}/tests/test_timestamp_utils.py +0 -0
  201. {osekit-1.3.0 → osekit-1.5.1}/tests/test_utils.py +0 -0
@@ -38,4 +38,4 @@ message: If you use this software, please cite it using the metadata from this f
38
38
  repository-code: https://github.com/Project-OSmOSE/OSEkit
39
39
  title: OSEkit
40
40
  type: software
41
- version: v1.3.0
41
+ version: v1.5.1
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.5
2
2
  Name: osekit
3
- Version: 1.3.0
3
+ Version: 1.5.1
4
4
  Summary: OSEkit
5
5
  Author-email: OSmOSE team <osmose@ensta-bretagne.fr>
6
6
  License-Expression: MIT
@@ -110,8 +110,6 @@ The data is fetched seamlessly on-demand from the audio file(s). The opening/clo
110
110
 
111
111
  Eventual time gap between audio items are filled with ``0.`` values.
112
112
 
113
- .. _normalization:
114
-
115
113
  Normalization
116
114
  """""""""""""
117
115
 
@@ -369,20 +367,20 @@ Once again, no audio has yet been fetched: everything happens only on-demand.
369
367
  NPZ matrices
370
368
  """"""""""""
371
369
 
372
- The ``SpectroData`` object can be used to compute the spectrum matrices of the ``AudioData`` with the :meth:`osekit.core.spectro_data.SpectroData.get_value` method.
370
+ The ``SpectroData`` object can be used to compute the STFT matrices of the ``AudioData`` with the :meth:`osekit.core.spectro_data.SpectroData.get_value` method.
373
371
 
374
- The :attr:`osekit.core.spectro_data.SpectroData.sx_dtype` property can be set to either ``complex`` (default) or ``float`` to return either the spectrum matrices as complex numbers or absolute values, respectively.
372
+ The :attr:`osekit.core.spectro_data.SpectroData.sx_dtype` property can be set to either ``complex`` (default) or ``float`` to return either the STFT matrices as complex numbers or absolute values, respectively.
375
373
 
376
- The spectrum matrices can be converted to decibels thanks to the :meth:`osekit.core.spectro_data.SpectroData.get_db_value` method.
377
- This method will convert the spectrum values either to dB SPL (re ``Instrument.P_REF``) if an :ref:`Instrument <instrument_calibration>` was provided to the ``AudioData`` or to dB FS otherwise.
374
+ The STFT matrices can be converted to decibels thanks to the :meth:`osekit.core.spectro_data.SpectroData.get_db_value` method.
375
+ This method will convert the STFT values either to dB SPL (re ``Instrument.P_REF``) if an :ref:`Instrument <instrument_calibration>` was provided to the ``AudioData`` or to dB FS otherwise.
378
376
 
379
- The spectrum matrices can then be exported to npz files thanks to the :meth:`osekit.core.spectro_data.SpectroData.write` method.
377
+ The STFT matrices can then be exported to npz files thanks to the :meth:`osekit.core.spectro_data.SpectroData.write` method.
380
378
 
381
379
  .. code-block:: python
382
380
 
383
381
  sd = SpectroData.from_audio_data(data=ad, fft=sft)
384
382
 
385
- sx_db = sd.get_db_value() # Use sd.get_value() if you want the complex spectrum
383
+ sx_db = sd.get_db_value() # Use sd.get_value() if you want the complex STFT matrix
386
384
 
387
385
  # If sx has already been computed, you can pass it as a parameter to avoid re-computing it.
388
386
  # Otherwise, it will just be computed from scratch.
@@ -463,7 +461,7 @@ LTAS Data
463
461
 
464
462
  OSEkit provides the :class:`osekit.core.ltas_data.LTASData` class for computing and plotting long-term spectral averages.
465
463
 
466
- When computing an STFT over a long audio recording, the resulting spectrum matrix can contain a very large number of time bins. ``LTASData`` reduces this temporal resolution by averaging STFT time bins while preserving the spectral information along the frequency axis.
464
+ When computing an STFT over a long audio recording, the resulting STFT matrix can contain a very large number of time bins. ``LTASData`` reduces this temporal resolution by averaging STFT time bins while preserving the spectral information along the frequency axis.
467
465
 
468
466
  The target number of time bins is specified with the ``n_bins`` parameter. OSEkit computes the result recursively so that the full STFT matrix never needs to be materialized when it would exceed this target size.
469
467
 
@@ -138,7 +138,7 @@
138
138
  "cell_type": "markdown",
139
139
  "id": "7bacba4e52a531c9",
140
140
  "metadata": {},
141
- "source": "Checking the required RAM to store the spectrum matrix can give a hint on when to switch to a LTAS computation:"
141
+ "source": "Checking the required RAM to store the STFT matrix can give a hint on when to switch to a LTAS computation:"
142
142
  },
143
143
  {
144
144
  "cell_type": "code",
@@ -144,8 +144,8 @@
144
144
  "\n",
145
145
  "transform = Transform(\n",
146
146
  " output_type=OutputType.SPECTROGRAM\n",
147
- " | OutputType.SPECTRUM, # we want to export both the spectrogram and the sx spectrum\n",
148
- " nb_ltas_time_bins=3000, # This will turn the regular spectrum computation in a LTAS\n",
147
+ " | OutputType.STFT_MATRIX, # we want to export both the spectrogram and the sx STFT matrix\n",
148
+ " nb_ltas_time_bins=3000, # This will turn the regular STFT matrix computation in a LTAS\n",
149
149
  " sample_rate=sample_rate,\n",
150
150
  " normalization=Normalization.DC_REJECT, # Removes the DC component\n",
151
151
  " fft=sft,\n",
@@ -173,7 +173,7 @@
173
173
  "cell_type": "markdown",
174
174
  "id": "8271f118422f38dc",
175
175
  "metadata": {},
176
- "source": "As for regular spectrum transforms, the output LTAS is stored in a `SpectroDataset` named after `transform.name`:"
176
+ "source": "As for regular STFT transforms, the output LTAS is stored in a `SpectroDataset` named after `transform.name`:"
177
177
  },
178
178
  {
179
179
  "cell_type": "code",
@@ -101,7 +101,7 @@
101
101
  "id": "1b00f1600c66e3c6",
102
102
  "metadata": {},
103
103
  "source": [
104
- "## Computing the spectrum of a specific channel\n",
104
+ "## Computing the spectrogram of a specific channel\n",
105
105
  "\n",
106
106
  "`SpectroData` target **a specific** channel of a file:"
107
107
  ]
@@ -144,7 +144,7 @@
144
144
  "from pandas import Timestamp, Timedelta\n",
145
145
  "\n",
146
146
  "transform = Transform(\n",
147
- " output_type=OutputType.SPECTRUM\n",
147
+ " output_type=OutputType.STFT_MATRIX\n",
148
148
  " | OutputType.SPECTROGRAM\n",
149
149
  " | OutputType.WELCH, # we want to export these three spectrum outputs\n",
150
150
  " begin=Timestamp(\"2022-09-25 22:35:15\"),\n",
@@ -99,7 +99,7 @@
99
99
  "cell_type": "markdown",
100
100
  "id": "8ae4db7363a92148",
101
101
  "metadata": {},
102
- "source": "Instantiate an **OSEkit** `SpectroData`, which is the object that will be used to plot spectrograms, export spectrum matrices etc."
102
+ "source": "Instantiate an **OSEkit** `SpectroData`, which is the object that will be used to plot spectrograms, export STFT matrices etc."
103
103
  },
104
104
  {
105
105
  "cell_type": "code",
@@ -133,7 +133,7 @@
133
133
  "source": [
134
134
  "import matplotlib.pyplot as plt\n",
135
135
  "\n",
136
- "# We store the spectrum matrix here to reuse it later in the notebook, this is not mandatory\n",
136
+ "# We store the STFT matrix here to reuse it later in the notebook, this is not mandatory\n",
137
137
  "sx = spectro_data.get_value()\n",
138
138
  "\n",
139
139
  "spectro_data.plot(sx=sx)\n",
@@ -182,7 +182,7 @@
182
182
  "cell_type": "markdown",
183
183
  "id": "93eae4edd8bb8bb5",
184
184
  "metadata": {},
185
- "source": "You can export both the PNG spectrogram and the NPZ spectrum matrix to file:"
185
+ "source": "You can export both the PNG spectrogram and the NPZ STFT matrix to file:"
186
186
  },
187
187
  {
188
188
  "cell_type": "code",
@@ -195,7 +195,7 @@
195
195
  },
196
196
  "outputs": [],
197
197
  "source": [
198
- "# Write a NPZ file with the spectrum matrix:\n",
198
+ "# Write a NPZ file with the STFT matrix:\n",
199
199
  "spectro_data.write(folder=audio_folder / \"export\", sx=sx)\n",
200
200
  "\n",
201
201
  "# Write a PNG image of the spectrogram\n",
@@ -34,7 +34,7 @@ In the ``docs/source/_static/sample_audio/timestamped`` folder, files are just n
34
34
 
35
35
  .. topic:: :doc:`Compute/plot a spectrogram <example_spectrogram>`
36
36
 
37
- | Compute the spectrum matrix and/or Power Spectral Density estimates of an ``AudioData``.
37
+ | Compute the STFT matrix and/or Power Spectral Density estimates of an ``AudioData``.
38
38
  | Export the matrices and/or plot a spectrogram.
39
39
 
40
40
  ===========
@@ -59,7 +59,7 @@ In the ``docs/source/_static/sample_audio/timestamped`` folder, files are just n
59
59
 
60
60
  .. topic:: :doc:`Work with multichannel audio files <example_multichannel>`
61
61
 
62
- Work on multichannel audio, compute spectrums from specific channels...
62
+ Work on multichannel audio, compute STFTs from specific channels...
63
63
 
64
64
  ===========
65
65
 
@@ -0,0 +1,27 @@
1
+ .. _job:
2
+
3
+ Job
4
+ ---
5
+
6
+ .. automodule:: osekit.job.job
7
+ :members: Job
8
+
9
+ Job Config
10
+ ----------
11
+
12
+ .. automodule:: osekit.job.config
13
+ :members: JobConfig
14
+
15
+ Job Scheduler
16
+ -------------
17
+
18
+ .. automodule:: osekit.job.scheduler.pbs
19
+ :members: Pbs
20
+ .. automodule:: osekit.job.scheduler.slurm
21
+ :members: Slurm
22
+
23
+ Job Builder
24
+ -----------
25
+
26
+ .. automodule:: osekit.job.builder
27
+ :members: JobBuilder
@@ -2,51 +2,65 @@ Working with jobs
2
2
  -----------------
3
3
 
4
4
  **OSEkit** can be set to send transform instructions to be computed on a remote server
5
- through the PBS queuing system.
5
+ through queuing systems.
6
6
 
7
7
  This feature has mainly be thought for the Public API, but it can nonetheless be used for
8
8
  any Core API operation.
9
9
 
10
- The job module is located at :mod:`osekit.utils.job`.
10
+ This is done thanks to the job package, located in :mod:`osekit.job`.
11
11
 
12
12
  Public API
13
13
  ^^^^^^^^^^
14
14
 
15
- Running Public API Analyses through PBS jobs only requires adding a :class:`osekit.utils.job.JobBuilder`
16
- instance to the :attr:`osekit.public.project.Project.job_builder` attribute:
15
+ Running Public API Analyses through jobs only requires adding a :class:`osekit.job.builder.JobBuilder`
16
+ instance to the :attr:`osekit.public.project.Project.job_builder` attribute.
17
+
18
+ The :class:`osekit.job.builder.JobBuilder`
19
+
20
+ Here is an example for running a transform on a PBS queue:
17
21
 
18
22
  .. code-block:: python
19
23
 
20
- from osekit.utils.job import JobConfig, JobBuilder
24
+ import os
25
+
26
+ from pandas import Timedelta
27
+
28
+ from osekit.job.builder import JobBuilder
29
+ from osekit.job.config import JobConfig
30
+ from osekit.job.scheduler.pbs import Pbs
21
31
  from osekit.public.project import Project
22
32
 
23
- project = Project(...) # See the Project documentation
33
+ project = Project(...) # See the Project documentation
24
34
 
25
35
  job_config = JobConfig(
26
- nb_nodes=1, # Number of nodes on which the job runs
27
- ncpus=28, # Number of total cores used per node
28
- ngpus=1, # Number of total GPU used per node
29
- mem="60gb", # Maximum amount of physical memory used by the job
30
- walltime=Timedelta(hours=5), # Maximum amount of real itime during which the job can be running
31
- venv_name=os.environ["CONDA_DEFAULT_ENV"], # Works only for conda venvs
32
- queue="omp" # Queue in which the job will be submitted
36
+ nb_nodes=1, # Number of nodes on which the job runs
37
+ ncpus=28, # Number of total cores used per node
38
+ ngpus=1, # Number of total GPU used per node
39
+ mem="60gb", # Maximum amount of physical memory used by the job
40
+ walltime=Timedelta(
41
+ hours=5
42
+ ), # Maximum amount of real itime during which the job can be running
43
+ venv_name=os.environ["CONDA_DEFAULT_ENV"], # Works only for conda venvs
33
44
  )
34
45
 
46
+ scheduler = Pbs(queue="omp") # Scheduler in which the job is submitted
47
+
35
48
  project.job_builder = JobBuilder(
36
49
  config=job_config,
50
+ scheduler=scheduler,
37
51
  )
38
52
 
39
53
  # Now the dataset has a non-None job_builder attribute,
40
54
  # running a transform will write a PBS file in the logs directory
41
- # and submit it to the requested queue.
55
+ # and submit it through the selected scheduler.
42
56
 
43
- project.run(...) # See the Transform documentation
57
+ project.run(...) # See the Transform documentation
44
58
 
45
59
 
46
60
  Core API
47
61
  ^^^^^^^^
48
62
 
49
- Exporting Core API datasets with jobs is doable by explicitly instantiating a :class:`osekit.utils.job.Job` object.
63
+ Exporting Core API datasets with jobs is doable by explicitly instantiating a :class:`osekit.job.job.Job` object.
50
64
 
51
65
  The export parameters are specified in the ``script_args`` parameter of the ``Job`` constructor,
52
66
  and follow the console arguments of the :mod:`osekit.public.export` script.
@@ -60,11 +74,13 @@ and follow the console arguments of the :mod:`osekit.public.export` script.
60
74
 
61
75
  from osekit.core.audio_dataset import AudioDataset
62
76
  from osekit.core.spectro_dataset import SpectroDataset
77
+ from osekit.job.config import JobConfig
78
+ from osekit.job.job import Job
79
+ from osekit.job.scheduler.pbs import Pbs
63
80
  from osekit.public import export_transform
64
81
 
65
82
  # Some Public API imports are required
66
83
  from osekit.public.transform import OutputType
67
- from osekit.utils.job import Job, JobConfig
68
84
 
69
85
  ads = AudioDataset(...) # See the AudioDataset doc
70
86
  sds = SpectroDataset(...) # See the SpectroDataset doc
@@ -109,9 +125,11 @@ and follow the console arguments of the :mod:`osekit.public.export` script.
109
125
  mem="60gb",
110
126
  walltime=Timedelta(hours=1),
111
127
  venv_name=os.environ["CONDA_DEFAULT_ENV"],
112
- queue="omp",
113
128
  )
114
129
 
130
+ # Scheduler configuration
131
+ scheduler = Pbs(queue="omp")
132
+
115
133
  job = Job(
116
134
  script_path=Path(export_transform.__file__),
117
135
  script_args=args,
@@ -120,12 +138,12 @@ and follow the console arguments of the :mod:`osekit.public.export` script.
120
138
  output_folder=Path(...), # Path in which the .out and .err files are written
121
139
  )
122
140
 
123
- # Write the PBS file and submit the job
124
- job.write_pbs(Path(...) / f"{job.name}.pbs")
125
- job.submit_pbs()
141
+ # Write the job file and submit it through the scheduler
142
+ scheduler.write(job=job, path=Path(...) / f"{job.name}.pbs")
143
+ scheduler.submit(job=job)
126
144
 
127
- You can then follow the status of the submitted job:
145
+ You can then follow the status of the submitted job through the scheduler:
128
146
 
129
147
  .. code-block:: python
130
148
 
131
- job.update_status()
149
+ scheduler.update_status(job=job)
@@ -37,10 +37,10 @@ Finally, a ``SpectroData`` that has a linked ``AudioData`` targets a **specific
37
37
  from scipy.signal import ShortTimeFFT
38
38
 
39
39
  sd = SpectroData.from_audio_data(data=ad, ...)
40
- sd.audio_channel = 2 # The spectrum will be computed on the channel 2 of the file
40
+ sd.audio_channel = 2 # The STFT will be computed on the channel 2 of the file
41
41
 
42
42
  .. important::
43
43
 
44
44
  The :attr:`osekit.core.spectro_data.SpectroData.audio_channel` value refers to the index of the channel of the **file**.
45
45
 
46
- If, as in the example above, the ``AudioFile`` has 3 channels ``[0,1,2]``, the ``AudioData`` targets channels ``[0,2]`` and the ``SpectroData`` targets the channel ``2``, the spectrum will be computed on the **third channel** of the file (with index ``2``).
46
+ If, as in the example above, the ``AudioFile`` has 3 channels ``[0,1,2]``, the ``AudioData`` targets channels ``[0,2]`` and the ``SpectroData`` targets the channel ``2``, the STFT will be computed on the **third channel** of the file (with index ``2``).
@@ -154,7 +154,7 @@ audio dataset (which could then be stored in another disk or partition):
154
154
  Running an ``Transform``
155
155
  ^^^^^^^^^^^^^^^^^^^^^^^^
156
156
 
157
- In **OSEkit**, **Analyses** are run with the :meth:`osekit.public.project.Project.run` method to process and export spectrogram images, spectrum matrices and audio files from original audio files.
157
+ In **OSEkit**, **Analyses** are run with the :meth:`osekit.public.project.Project.run` method to process and export spectrogram images, STFT matrices and audio files from original audio files.
158
158
 
159
159
  .. note::
160
160
 
@@ -175,8 +175,8 @@ The ``output_type`` parameter passed to the initializer is a :class:`osekit.publ
175
175
  - Output
176
176
  * - ``OutputType.AUDIO``
177
177
  - Reshaped audio files
178
- * - ``OutputType.SPECTRUM``
179
- - `STFT <https://docs.scipy.org/doc/scipy/reference/generated/scipy.signal.ShortTimeFFT.stft.html#scipy.signal.ShortTimeFFT.stft>`_ NPZ spectrum files
178
+ * - ``OutputType.STFT_MATRIX``
179
+ - `STFT <https://docs.scipy.org/doc/scipy/reference/generated/scipy.signal.ShortTimeFFT.stft.html#scipy.signal.ShortTimeFFT.stft>`_ NPZ STFT files
180
180
  * - ``OutputType.WELCH``
181
181
  - `Welch <https://docs.scipy.org/doc/scipy/reference/generated/scipy.signal.welch.html>`_ NPZ files
182
182
  * - ``OutputType.SPECTROGRAM``
@@ -199,7 +199,7 @@ The remaining parameters of the transform (begin and end **Timestamps**, duratio
199
199
 
200
200
  .. note::
201
201
 
202
- If the ``Transform`` contains spectral computations (either ``OutputType.SPECTRUM``, ``OutputType.SPECTROGRAM`` or ``OutputType.WELCH`` is in ``output_type``), a `scipy ShortTimeFFT instance <https://docs.scipy.org/doc/scipy/reference/generated/scipy.signal.ShortTimeFFT.html#scipy.signal.ShortTimeFFT>`_ should be passed to the ``Transform`` initializer.
202
+ If the ``Transform`` contains spectral computations (either ``OutputType.STFT_MATRIX``, ``OutputType.SPECTROGRAM`` or ``OutputType.WELCH`` is in ``output_type``), a `scipy ShortTimeFFT instance <https://docs.scipy.org/doc/scipy/reference/generated/scipy.signal.ShortTimeFFT.html#scipy.signal.ShortTimeFFT>`_ should be passed to the ``Transform`` initializer.
203
203
 
204
204
  .. _editing_transform:
205
205
 
@@ -321,7 +321,7 @@ The ``cool_reshape`` folder has been created, containing the freshly created ``1
321
321
  Example: full transform
322
322
  """""""""""""""""""""""
323
323
 
324
- Let's now say we want to export audio, spectrum matrices and spectrograms with the following parameters:
324
+ Let's now say we want to export audio, STFT matrices and spectrograms with the following parameters:
325
325
 
326
326
  .. list-table:: Example transform parameters
327
327
  :widths: 40 60
@@ -367,9 +367,9 @@ Then we are all set for running the transform:
367
367
 
368
368
  transform = Transform(
369
369
  output_type=OutputType.AUDIO
370
- | OutputType.SPECTRUM
370
+ | OutputType.STFT_MATRIX
371
371
  | OutputType.WELCH
372
- | OutputType.SPECTROGRAM, # Full transform : audio files, spectrum matrices and spectrograms will be exported.
372
+ | OutputType.SPECTROGRAM, # Full transform : audio files, STFT matrices and spectrograms will be exported.
373
373
  begin=project.origin_dataset.begin
374
374
  + Timedelta(minutes=30), # 30m after the begin of the original dataset
375
375
  end=project.origin_dataset.begin
@@ -1,6 +1,6 @@
1
1
  [project]
2
2
  name = "osekit"
3
- version = "1.3.0"
3
+ version = "1.5.1"
4
4
  description = "OSEkit"
5
5
  readme = "README.md"
6
6
  license = "MIT"
@@ -135,18 +135,13 @@ class AudioData(BaseData[AudioItem, AudioFile]):
135
135
 
136
136
  @property
137
137
  def channels(self) -> list[int]:
138
- """The Butterworth filter to apply to the audio data."""
139
- return self._channels
138
+ """The channels targeted by this ``AudioData``."""
139
+ return list({channel for i in self.items for channel in i.channels})
140
140
 
141
141
  @channels.setter
142
142
  def channels(self, value: list[int] | None) -> None:
143
- if value is None:
144
- nb_channels_max = max(
145
- [1]
146
- + [item.nb_channels for item in self.items if type(item) is AudioItem],
147
- )
148
- value = list(range(nb_channels_max))
149
- self._channels = value
143
+ for item in self.items:
144
+ item.channels = value
150
145
 
151
146
  @classmethod
152
147
  def _make_item(
@@ -261,16 +256,20 @@ class AudioData(BaseData[AudioItem, AudioFile]):
261
256
  else self.butter.filter(sig=output, fs=self.sample_rate)
262
257
  )
263
258
 
264
- @staticmethod
265
259
  def _flush(
260
+ self: AudioData,
266
261
  resampler: soxr.ResampleStream,
267
262
  remaining_samples: int,
268
263
  ) -> np.ndarray:
269
- flush = resampler.resample_chunk(np.array([]), last=True)
264
+ empty_values = np.empty((0, self.nb_channels))
265
+ flush = resampler.resample_chunk(
266
+ empty_values,
267
+ last=True,
268
+ )
270
269
  if len(flush) == 0:
271
- return np.array([])[:, None]
270
+ return empty_values
272
271
  if not remaining_samples:
273
- return np.array([])[:, None]
272
+ return empty_values
274
273
  flush = flush[:remaining_samples]
275
274
  return flush[:, None] if flush.ndim == 1 else flush
276
275
 
@@ -297,9 +296,16 @@ class AudioData(BaseData[AudioItem, AudioFile]):
297
296
  for item in self.items:
298
297
  if item.is_empty:
299
298
  silence_length = round(item.duration.total_seconds() * self.sample_rate)
300
- yield item.get_value().repeat(
301
- silence_length,
302
- axis=0,
299
+ yield (
300
+ item.get_value()
301
+ .repeat(
302
+ silence_length,
303
+ axis=0,
304
+ )
305
+ .repeat(
306
+ self.nb_channels,
307
+ axis=1,
308
+ )
303
309
  )
304
310
  produced_samples += silence_length
305
311
  continue
@@ -326,7 +332,8 @@ class AudioData(BaseData[AudioItem, AudioFile]):
326
332
  )
327
333
 
328
334
  for chunk in item.stream(chunk_size=chunk_size):
329
- y = chunk[:, self.channels]
335
+ y = np.zeros((chunk.shape[0], self.nb_channels))
336
+ y[:, self._get_item_channel_mapping(item=item)] = chunk
330
337
  if item.sample_rate != self.sample_rate:
331
338
  y = resampler.resample_chunk(x=chunk)
332
339
 
@@ -347,6 +354,15 @@ class AudioData(BaseData[AudioItem, AudioFile]):
347
354
  remaining_samples=total_samples - produced_samples,
348
355
  )
349
356
 
357
+ def _get_item_channel_mapping(self, item: AudioItem) -> list[int]:
358
+ """Return the mapping of the item channels in the AudioData channels.
359
+
360
+ If the AudioItem targets channels [2,4] and the AudioData target channels
361
+ [0,2,3,4], the channels with indexes 0 and 1 of the item data (channels 2 and 4) must be mapped
362
+ to the channels with indexes 1 and 3 of the AudioData (channels 2 and 4).
363
+ """
364
+ return [self.channels.index(channel) for channel in item.channels]
365
+
350
366
  def get_value(self) -> np.ndarray:
351
367
  """Return the value of the audio data.
352
368
 
@@ -22,6 +22,7 @@ class AudioItem(BaseItem[AudioFile]):
22
22
  file: AudioFile | None = None,
23
23
  begin: Timestamp | None = None,
24
24
  end: Timestamp | None = None,
25
+ channels: list[int] | None = None,
25
26
  ) -> None:
26
27
  """Initialize an ``AudioItem`` from an ``AudioFile`` and begin/end timestamps.
27
28
 
@@ -35,9 +36,12 @@ class AudioItem(BaseItem[AudioFile]):
35
36
  end: pandas.Timestamp (optional)
36
37
  The timestamp at which this item ends.
37
38
  It is defaulted to ``file.end``.
39
+ channels: list[int]
40
+ Considered channels of the linked audio file.
38
41
 
39
42
  """
40
43
  super().__init__(file=file, begin=begin, end=end)
44
+ self.channels = channels
41
45
 
42
46
  @property
43
47
  def sample_rate(self) -> float:
@@ -46,8 +50,25 @@ class AudioItem(BaseItem[AudioFile]):
46
50
 
47
51
  @property
48
52
  def nb_channels(self) -> int:
49
- """Number of channels in the associated ``AudioFile``."""
50
- return 1 if self.is_empty else self.file.channels
53
+ """Number of channels targeted in the associated ``AudioFile``."""
54
+ return len(self.channels)
55
+
56
+ @property
57
+ def channels(self) -> list[int]:
58
+ """The channels targeted by this ``AudioItem``."""
59
+ return self._channels
60
+
61
+ @channels.setter
62
+ def channels(self, value: list[int] | None) -> None:
63
+ if self.is_empty:
64
+ self._channels = [0]
65
+ return
66
+
67
+ if value is None:
68
+ self._channels = list(range(self.file.channels))
69
+ return
70
+
71
+ self._channels = [channel for channel in value if channel < self.file.channels]
51
72
 
52
73
  @property
53
74
  def shape(self) -> tuple[int, int]:
@@ -78,5 +99,5 @@ class AudioItem(BaseItem[AudioFile]):
78
99
 
79
100
  while remaining > 0:
80
101
  frames_to_read = min(chunk_size, remaining)
81
- yield self.file.stream(chunk_size=frames_to_read)
102
+ yield self.file.stream(chunk_size=frames_to_read)[:, self.channels]
82
103
  remaining -= frames_to_read
@@ -330,15 +330,16 @@ class BaseDataset[TData: BaseData, TFile: BaseFile](Event, ABC):
330
330
  End of the last data object.
331
331
  Defaulted to the end of the last file.
332
332
  mode: Literal["files", "timedelta_total", "timedelta_file"]
333
- Mode of creation of the dataset data from the original files.
333
+ Mode of creation of the dataset data from the files.
334
334
  ``"files"``: one data will be created for each file.
335
335
  ``"timedelta_total"``: data objects of duration equal to ``data_duration`` will
336
336
  be created from the ``begin`` timestamp to the ``end`` timestamp.
337
337
  ``"timedelta_file"``: data objects of duration equal to ``data_duration`` will
338
- be created from the beginning of the first file that the ``begin`` timestamp
339
- is into, until it would resume in a data beginning between two files.
338
+ be created from the ``begin`` timestamp, until it would resume in a data
339
+ beginning between two files.
340
340
  Then, the next data object will be created from the
341
- beginning of the next original file and so on.
341
+ beginning of the next file and so on until the ``end`` timestamp (if
342
+ specified) or the last file's end is reached.
342
343
  data_duration: Timedelta | None
343
344
  Duration of the data objects.
344
345
  If mode is set to ``"files"``, this parameter has no effect.
@@ -481,14 +482,16 @@ class BaseDataset[TData: BaseData, TFile: BaseFile](Event, ABC):
481
482
  continue
482
483
  files_chunk = [file]
483
484
 
485
+ next_jump = None
484
486
  for next_file in files[idx + 1 :]:
485
487
  upper_data_limit = last_window_end(
486
- begin=file.begin,
488
+ begin=begin,
487
489
  end=files_chunk[-1].end,
488
490
  window_hop=data_hop,
489
491
  window_duration=data_duration,
490
492
  )
491
493
  if upper_data_limit < next_file.begin:
494
+ next_jump = next_file
492
495
  break
493
496
  files_chunk.append(next_file)
494
497
 
@@ -500,13 +503,16 @@ class BaseDataset[TData: BaseData, TFile: BaseFile](Event, ABC):
500
503
  **kwargs,
501
504
  )
502
505
  for data_begin in date_range(
503
- file.begin,
504
- files_chunk[-1].end,
506
+ begin,
507
+ min(files_chunk[-1].end, end),
505
508
  freq=data_hop,
506
509
  inclusive="left",
507
510
  )
508
511
  )
509
512
 
513
+ if next_jump:
514
+ begin = next_jump.begin
515
+
510
516
  return output
511
517
 
512
518
  @classmethod