smftools 0.2.1__py3-none-any.whl → 0.2.3__py3-none-any.whl

{smftools-0.2.1.dist-info → smftools-0.2.3.dist-info}/RECORD

@@ -1,15 +1,21 @@
-smftools/__init__.py,sha256=OXW2_b5NUGZhTXsH8qY0PzfJnaz8T2y6lCqMnSVSuIk,676
+smftools/__init__.py,sha256=aZlrZBVexf_nEnzQeZu7NU_Kp6OnxcYpLo1KPImi7sI,599
 smftools/_settings.py,sha256=Ed8lzKUA5ncq5ZRfSp0t6_rphEEjMxts6guttwTZP5Y,409
-smftools/_version.py,sha256=tC9CwL4Nm8brVXJnZNGk_eoZaJj6eOtLKtOrdJMrpoI,21
-smftools/cli.py,sha256=MNObu279y322JHkmugssM0rVHo0UQ1zboTG9MlqnMgQ,7033
-smftools/load_adata.py,sha256=VJMUBqRC8InIj48JMnkZKLuqEz1u8uSTNx_ARl0cn7M,74313
-smftools/readwrite.py,sha256=ObNxBj6Y_zIHqQpAvmHAddAypLjg7F3qARF-sH-V3do,42706
+smftools/_version.py,sha256=X0PliCRFAeVnSTceUeHX1eM0j1HFhGFDWCRxLdde2Bs,21
+smftools/cli_entry.py,sha256=_QdtEKcVK5o-e5s9ETB9sOIdftPVlrDxvvjBKcP6YNk,14680
+smftools/readwrite.py,sha256=ExKZHNZ0QB-PtSck08drXfHTqbPeSUTHiYhv951SH1s,45994
+smftools/cli/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+smftools/cli/cli_flows.py,sha256=xRiFUThoAL3LX1xdXaHVg4LjyJI4uNpGsc9aQ_wVCto,4941
+smftools/cli/hmm_adata.py,sha256=PApUJW0lO4kcLjsiqqQopXgL3Dg-AascIqJrgvSY1Rg,15916
+smftools/cli/load_adata.py,sha256=Qt1ej-osyJ47fpBkGaSDgR1F8E4aBNAdcXeBAGM-Lqg,29100
+smftools/cli/preprocess_adata.py,sha256=EKGbSTli7qvL44OQUmMalYJjsH9vn3w4Rx7U7BL0ybs,20991
+smftools/cli/spatial_adata.py,sha256=AX6iyBfbXud9actteTvDuaQUU_SE3SyBIeknR317g34,30212
 smftools/config/__init__.py,sha256=ObUnnR7aRSoD_uvpmsxA_BUFt4NOOfWNopDVCqjp7tg,69
-smftools/config/conversion.yaml,sha256=rJGhrVd95p6_6OVxLq2lvobJu8SGzNYI80jU0fLeK_g,795
-smftools/config/deaminase.yaml,sha256=Vh3Wg0bCb88S20Ob-8zi3eQJ1g_pcBulR9pPbAX9U1o,1138
-smftools/config/default.yaml,sha256=0DYIvvdbzoB2eJgsoxEzx4Rc0TVGaiHa85nxo1VwCqQ,9704
-smftools/config/direct.yaml,sha256=2F_fGploWW3f88Y7sTZ68Vk9fgNaO-sb5AK-Cutc2TQ,735
-smftools/config/experiment_config.py,sha256=zQhWaag9hPuexnTOqZ-Od--c3iHs18c4Wc2sU-LOyts,52872
+smftools/config/conversion.yaml,sha256=HrFz2f9QRe1RuhmgU6ZtMHaM4ZzY61_aLcugsmpV40Q,969
+smftools/config/deaminase.yaml,sha256=mw2aY222y2xg08Rs5CWvjlrXo3vaEim7JwBThA80y4o,1349
+smftools/config/default.yaml,sha256=3IrX0OrUyjhVc3CqTjM8uiprKWrrBdVtil4YhtVzKdQ,10233
+smftools/config/direct.yaml,sha256=SBhdtG7PKm-z5xxQmA7JV3NQsGnUJ4p58fGH8BnoMrM,2137
+smftools/config/discover_input_files.py,sha256=G9vyAmK_n_8Ur5dOnumevVLG3ydHchMy_JQrJdiuuz0,3892
+smftools/config/experiment_config.py,sha256=d_6f_Uv3CY-1orHbxpHtAZDsY2gwxw079_pNgR9wDUg,58837
 smftools/datasets/F1_hybrid_NKG2A_enhander_promoter_GpC_conversion_SMF.h5ad.gz,sha256=q6wJtgFRDln0o20XNCx1qad3lwcdCoylqPN7wskTfI8,2926497
 smftools/datasets/F1_sample_sheet.csv,sha256=9PodIIOXK2eamYPbC6DGnXdzgi9bRDovf296j1aM0ak,259
 smftools/datasets/__init__.py,sha256=xkSTlPuakVYVCuRurif9BceNBDt6bsngJvvjI8757QI,142
@@ -24,58 +30,63 @@ smftools/hmm/display_hmm.py,sha256=3WuQCPvM3wPfzAdgbhfiBTd0g5mQdx9HTUdqAxs2aj4,8
 smftools/hmm/hmm_readwrite.py,sha256=DjJ3hunpBQ7N0GVvxL7-0QUas_SkA88LVgL72mVK2cI,359
 smftools/hmm/nucleosome_hmm_refinement.py,sha256=nQWimvse6dclcXhbU707rGbRVMKHM0mU_ZhH9g2yCMA,4641
 smftools/hmm/train_hmm.py,sha256=srzRcB9LEmNuHyBM0R5Z0VEnxecifQt-MoaJhADxGT8,2477
-smftools/informatics/__init__.py,sha256=8tvVG08L_Z-bP28PusBtVt1UTnHxuKi0lImLNcP7qso,338
-smftools/informatics/basecall_pod5s.py,sha256=Ynmxscsxj6qp-zVY0RWodq513oDuHDaHnpqoepB3RUU,3930
-smftools/informatics/fast5_to_pod5.py,sha256=h-cUZX5sWwPCkQ4g3kyz3koSBjZOWI6EjSpWO8zib1I,862
-smftools/informatics/readwrite.py,sha256=DgVisHYdkjzaO7suPbUvluImeTc3jqGDlioNveHUxPc,4158
-smftools/informatics/subsample_fasta_from_bed.py,sha256=YqYV09rvEQdeiS5hTTrKa8xYmJfeM3Vk-UUqwpw0qBk,1983
-smftools/informatics/subsample_pod5.py,sha256=zDw9tRcrFRmPI62xkcy9dh8IfsJcuYm7R-FVeBC_g3s,4701
+smftools/informatics/__init__.py,sha256=vLvSrCtCVYRUCCNLW7fL3ltPr3h_w8FhT--V6el3ZkQ,1191
+smftools/informatics/bam_functions.py,sha256=otgl3TRPLn5Fnsx1jXX75du90k3XB3RHGzlfamvETsU,32670
+smftools/informatics/basecalling.py,sha256=jc39jneaa8Gt1azutHgBGWHqCoPeTVSGBu3kyQwP7xM,3460
+smftools/informatics/bed_functions.py,sha256=uETVxT5mRWDNn7t0OqhDi8kDiq7uDakeHB1L2JsP4PA,13377
+smftools/informatics/binarize_converted_base_identities.py,sha256=yOepGaNBGfZJEsMiLRwKauvsmaHn_JRrxaGp8LmKAXs,7778
+smftools/informatics/complement_base_list.py,sha256=k6EkLtxFoajaIufxw1p0pShJ2nPHyGLTbzZmIFFjB4o,532
+smftools/informatics/converted_BAM_to_adata.py,sha256=Y2kQNWly0WjjGN9El9zL1nLfjVxmPLWONvX5VNgZUh0,22554
+smftools/informatics/fasta_functions.py,sha256=5IfTkX_GIj5gRJB9PjL_WjyEktpBHwGsmS_nnO1ETjI,9790
+smftools/informatics/h5ad_functions.py,sha256=iAOxJjhaDslTUC78kjUHlCELigDl73sWo0fvXcKuFoI,7824
+smftools/informatics/modkit_extract_to_adata.py,sha256=TrgrL_IgfqzNJ9qZ_2EvF_B38_Syw8mP38Sl7v0Riwo,55278
+smftools/informatics/modkit_functions.py,sha256=lywjeqAJ7Cdd7k-0P3YaL_9cAZvEDTDLh91rIRcSMWE,5604
+smftools/informatics/ohe.py,sha256=MEmh3ps-ZSSyXuIrr5LMzQvCsDJRCYiy7JS-WD4TlYs,5805
+smftools/informatics/pod5_functions.py,sha256=vxwhD_d_iWpJydIpbf0uce7VGHm8sBnCwb7tLNpYBc8,9859
+smftools/informatics/run_multiqc.py,sha256=n6LvQuGQpLfsutVGmgvHfV0SV5PqTQ8wa_SeKOjRssM,1052
 smftools/informatics/archived/bam_conversion.py,sha256=I8EzXjQixMmqx2oWnoNSH5NURBhfT-krbWHkoi_M964,3330
 smftools/informatics/archived/bam_direct.py,sha256=jbEFtUIiUR8Wlp3po_sWkr19AUNS9WZjglojb9j28vo,3606
+smftools/informatics/archived/basecall_pod5s.py,sha256=Ynmxscsxj6qp-zVY0RWodq513oDuHDaHnpqoepB3RUU,3930
 smftools/informatics/archived/basecalls_to_adata.py,sha256=-Nag6lr_NAtU4t8jo0GSMdgIAIfmDge-5VEUPQbEatE,3692
 smftools/informatics/archived/conversion_smf.py,sha256=QhlISVi3Z-XqFKyDG_CenLojovAt5-ZhuVe9hus36lg,7177
 smftools/informatics/archived/deaminase_smf.py,sha256=mNeg1mIYYVLIiW8powEpz0CqrGRDsrmY5-aoIgwMGHs,7221
 smftools/informatics/archived/direct_smf.py,sha256=ylPGFBvRLdxLHeDJjAwq98j8Q8_lfGK3k5JJnQxrwJw,7485
+smftools/informatics/archived/fast5_to_pod5.py,sha256=TRG_FYYGCGWUPzZCt0ZqzB8gQv_HKvkssp9nTctWzXU,1398
 smftools/informatics/archived/print_bam_query_seq.py,sha256=8Z2ZJEOOlfWYUXiZGjteLWU4yTgvV8KQzEIBHUmamGM,838
-smftools/informatics/helpers/__init__.py,sha256=EgCIcJ6o3_R3vzsFwhtvOcKKWnmmMmN_GZXDQ_K_-NI,2693
-smftools/informatics/helpers/align_and_sort_BAM.py,sha256=gy_BU6KfDd584LPFybJ7JzNwfCD95dZXx6MccnT4Qro,3725
-smftools/informatics/helpers/aligned_BAM_to_bed.py,sha256=e6yg5-yHcw0QPFI3oRVHrhfAUj7US77Ir2VVzE3c-x8,3374
-smftools/informatics/helpers/bam_qc.py,sha256=IlrXXpCdTYIv_89SE8D5tJ1wtTzxWGjk9vc-rbC1UjU,2430
-smftools/informatics/helpers/bed_to_bigwig.py,sha256=AazYEZzKgKgukSFwCpeiApzxh1kbt11X4RFqRIiBIaY,1466
-smftools/informatics/helpers/binarize_converted_base_identities.py,sha256=yOepGaNBGfZJEsMiLRwKauvsmaHn_JRrxaGp8LmKAXs,7778
-smftools/informatics/helpers/canoncall.py,sha256=5WS6lwukc_xYTdPQy0OSj-WLbx0Rg70Cun1lCucY7w8,1741
-smftools/informatics/helpers/complement_base_list.py,sha256=k6EkLtxFoajaIufxw1p0pShJ2nPHyGLTbzZmIFFjB4o,532
-smftools/informatics/helpers/concatenate_fastqs_to_bam.py,sha256=0jy4H1ORuqaarsznv9tS1SM8CCRjaaD20NMknNvQPv0,16212
-smftools/informatics/helpers/converted_BAM_to_adata.py,sha256=sRmOtn0kNosLYfogqslDHg1Azk51l6nfNOLgQOnQjlA,14591
-smftools/informatics/helpers/converted_BAM_to_adata_II.py,sha256=9Tz-qWtK9v1DTlK6yManvhOlMcaHFQUmmrRZQ5eiECw,22229
-smftools/informatics/helpers/count_aligned_reads.py,sha256=uYyUYglF1asiaoxr-LKxPMUEbfyD7FS-dumTg2hJHzQ,2170
-smftools/informatics/helpers/demux_and_index_BAM.py,sha256=2B_UiU05ln3gYvcN9aC_w6qs8j_WAF4pHWZekAYsXm4,2114
-smftools/informatics/helpers/discover_input_files.py,sha256=hUes2iKBQW_sVmAYD-1JnLD9Ub-COEHzrRKWNFipl0g,3725
-smftools/informatics/helpers/extract_base_identities.py,sha256=2yvr5uff9ah0jylFjNMt7oRJb1z_YdhvM6htSxI0frg,3038
-smftools/informatics/helpers/extract_mods.py,sha256=MbSIiyj3zx7WlSSWMRPriLMkBtxYc1EWZiAAirMVgqA,3865
-smftools/informatics/helpers/extract_read_features_from_bam.py,sha256=SYAb4Q1HxiJzCx5bIz86MdH_TvVPsRAVodZD9082HGY,1491
-smftools/informatics/helpers/extract_read_lengths_from_bed.py,sha256=Cw39wgp1eBTV45Wk1l0c9l-upBW5N2OcgyWXTAXln90,678
-smftools/informatics/helpers/extract_readnames_from_BAM.py,sha256=3FxSNqbZ1VsOK2RfHrvevQTzhWATf5E8bZ5yVOqayvk,759
-smftools/informatics/helpers/find_conversion_sites.py,sha256=JPlDipmzeCBkV_T6esGD5ptwmbQmk8gJMTh7NMaSYd4,2480
-smftools/informatics/helpers/generate_converted_FASTA.py,sha256=UniQfERNt4FC5L8T1tzr4cLQOJc3wMBPhuWmC-lC8Fs,3747
-smftools/informatics/helpers/get_chromosome_lengths.py,sha256=sLumLrGsU_Xg_oJcdOpQyjUGpJoT2HbcmxWwbwzXUlE,1036
-smftools/informatics/helpers/get_native_references.py,sha256=fRuyEm9UJkfd5DwHmFb1bxEtNvtSI1_BxGRmrCymGkw,981
-smftools/informatics/helpers/index_fasta.py,sha256=N3IErfSiavYldeaat8xcQgA1MpykoQHcE0gHUeWuClE,267
-smftools/informatics/helpers/make_dirs.py,sha256=lWHXpwC76MFM5sSme9i_WeYUaxutzybendokhny03ds,537
-smftools/informatics/helpers/make_modbed.py,sha256=cOQ97gPfRiCcw_fqboxousXIiOYjp78IFYLbu749U1Y,939
-smftools/informatics/helpers/modQC.py,sha256=LeOBObG8gAVVdgESIMceYhd5AW1gfN7ABo91OQtOzTM,1041
-smftools/informatics/helpers/modcall.py,sha256=LVPrdMNVp2gyQTJ4BNp8NJNm89AueDjsKaY7Gqkluho,1777
-smftools/informatics/helpers/modkit_extract_to_adata.py,sha256=yjG_Onh6YgwpE11I8mgQyN6F-8yStJgvDcI38L13i4A,52098
-smftools/informatics/helpers/ohe_batching.py,sha256=QVOiyl9fYHNIFWM23afYnQo0uaOjf1NR3ASKGVSrmuw,2975
-smftools/informatics/helpers/ohe_layers_decode.py,sha256=gIgUC9L8TFLi-fTnjR4PRzXdUaH5D6WL2Hump6XOoy0,1042
-smftools/informatics/helpers/one_hot_decode.py,sha256=3n4rzY8_aC9YKmgrftsguMsH7fUyQ-DbWmrOYF6la9s,906
-smftools/informatics/helpers/one_hot_encode.py,sha256=5hHigA6-SZLK84WH_RHo06F_6aTg7S3TJgvSr8gxGX8,1968
-smftools/informatics/helpers/plot_bed_histograms.py,sha256=sdtz_ieU_5rz8WyfAzjxbzY_w8kLdE_Rklvjax1hl3Q,10442
-smftools/informatics/helpers/run_multiqc.py,sha256=qkw48DeBdTEqzhKFGjMUlvNmTehp8wRPkcxdkwERkHc,980
-smftools/informatics/helpers/separate_bam_by_bc.py,sha256=WJZwKCYODUvzFaVWwX3SUE8sxEXmeYmSi7Dl9h2J2EY,1802
-smftools/informatics/helpers/split_and_index_BAM.py,sha256=yowMusTGoC7uRD0jAwOHzBegX6MV7f-uY-XSzkX5cBw,1253
-smftools/informatics/helpers/archived/informatics.py,sha256=gKb2ZJ_LcAeEXuQqn9e-QDF_sS4tMpMTr2vZlqa7n54,14572
-smftools/informatics/helpers/archived/load_adata.py,sha256=DhvYYqO9VLsZqhL1WjN9sd-e3fgvdXGlgTP18z1h0L0,33654
+smftools/informatics/archived/subsample_fasta_from_bed.py,sha256=7YTKhXg_mtP4KWpnD-TB4nuFEL4crOa9_d84IJKllyQ,1633
+smftools/informatics/archived/subsample_pod5.py,sha256=zDw9tRcrFRmPI62xkcy9dh8IfsJcuYm7R-FVeBC_g3s,4701
+smftools/informatics/archived/helpers/archived/__init__.py,sha256=DiiBerFJAxZeG5y0ScpJSaVBJ8b4XWdfEJCh8Q7k8jU,2783
+smftools/informatics/archived/helpers/archived/align_and_sort_BAM.py,sha256=yaRfhQDh3HpsSTme6QnSqBgElCS0kv2G6TunhvR1weY,5493
+smftools/informatics/archived/helpers/archived/aligned_BAM_to_bed.py,sha256=N3NAOaoSt_M4V48vtTP_m_iF1tRuNIPS_uNJ3Y0IA4E,3391
+smftools/informatics/archived/helpers/archived/bam_qc.py,sha256=PWl3dViCHGOcjB4UKkxBFz34Gc0PXHVTHjpYVNckVH0,7975
+smftools/informatics/archived/helpers/archived/bed_to_bigwig.py,sha256=Bg9wFsavUU9Ha57n_99vYlYpVcbDUz3tLtYJ7ZFVR9k,2986
+smftools/informatics/archived/helpers/archived/canoncall.py,sha256=5WS6lwukc_xYTdPQy0OSj-WLbx0Rg70Cun1lCucY7w8,1741
+smftools/informatics/archived/helpers/archived/concatenate_fastqs_to_bam.py,sha256=6GTHXG1dfaC8rBin5NthG3xgyGqOsT6wIGxJVCmCq58,9774
+smftools/informatics/archived/helpers/archived/converted_BAM_to_adata.py,sha256=sRmOtn0kNosLYfogqslDHg1Azk51l6nfNOLgQOnQjlA,14591
+smftools/informatics/archived/helpers/archived/count_aligned_reads.py,sha256=ZF_kkzAf1RvM4PwDYhxD36UiuVuMM_MBvZgiXom1NQ0,2176
+smftools/informatics/archived/helpers/archived/demux_and_index_BAM.py,sha256=KmU7nqGQ-MfDrp8h3txbToGn4h95Rkvg0WEiuext-vY,2000
+smftools/informatics/archived/helpers/archived/extract_base_identities.py,sha256=CaFqNBjkDujYlyiUnOeRock1OQWs3CeiD3yTL96sjIs,3043
+smftools/informatics/archived/helpers/archived/extract_mods.py,sha256=Mrs7mrLFgCTiRGfPFSyvJm6brq--LGzZrNDiFB-jynI,3895
+smftools/informatics/archived/helpers/archived/extract_read_features_from_bam.py,sha256=SYAb4Q1HxiJzCx5bIz86MdH_TvVPsRAVodZD9082HGY,1491
+smftools/informatics/archived/helpers/archived/extract_read_lengths_from_bed.py,sha256=Cw39wgp1eBTV45Wk1l0c9l-upBW5N2OcgyWXTAXln90,678
+smftools/informatics/archived/helpers/archived/extract_readnames_from_BAM.py,sha256=3FxSNqbZ1VsOK2RfHrvevQTzhWATf5E8bZ5yVOqayvk,759
+smftools/informatics/archived/helpers/archived/find_conversion_sites.py,sha256=JPlDipmzeCBkV_T6esGD5ptwmbQmk8gJMTh7NMaSYd4,2480
+smftools/informatics/archived/helpers/archived/generate_converted_FASTA.py,sha256=Us6iH1cIhsXDnTvDxI-FEHB6ndbB30hd1ss-9dIoWVE,3819
+smftools/informatics/archived/helpers/archived/get_chromosome_lengths.py,sha256=BEroXshYSpjf5wt_vrEAFiTJmSuf-kvD-Z1B_1gusME,1000
+smftools/informatics/archived/helpers/archived/get_native_references.py,sha256=fRuyEm9UJkfd5DwHmFb1bxEtNvtSI1_BxGRmrCymGkw,981
+smftools/informatics/archived/helpers/archived/index_fasta.py,sha256=w6xHFSaoXVk-YWZWftZ9Xv8rywZ_IuuIouLQ12KL3ro,779
+smftools/informatics/archived/helpers/archived/informatics.py,sha256=gKb2ZJ_LcAeEXuQqn9e-QDF_sS4tMpMTr2vZlqa7n54,14572
+smftools/informatics/archived/helpers/archived/load_adata.py,sha256=DhvYYqO9VLsZqhL1WjN9sd-e3fgvdXGlgTP18z1h0L0,33654
+smftools/informatics/archived/helpers/archived/make_modbed.py,sha256=Wh0UCSOL4fMZbWYK-3oGGHwJtqPurJ3Bl6wJWBaTXoM,923
+smftools/informatics/archived/helpers/archived/modQC.py,sha256=pz2EscFgO-j-9dfNgNDseweXXqM5-a-Rj2abBLErLd0,1051
+smftools/informatics/archived/helpers/archived/modcall.py,sha256=LVPrdMNVp2gyQTJ4BNp8NJNm89AueDjsKaY7Gqkluho,1777
+smftools/informatics/archived/helpers/archived/ohe_batching.py,sha256=QVOiyl9fYHNIFWM23afYnQo0uaOjf1NR3ASKGVSrmuw,2975
+smftools/informatics/archived/helpers/archived/ohe_layers_decode.py,sha256=gIgUC9L8TFLi-fTnjR4PRzXdUaH5D6WL2Hump6XOoy0,1042
+smftools/informatics/archived/helpers/archived/one_hot_decode.py,sha256=3n4rzY8_aC9YKmgrftsguMsH7fUyQ-DbWmrOYF6la9s,906
+smftools/informatics/archived/helpers/archived/one_hot_encode.py,sha256=5hHigA6-SZLK84WH_RHo06F_6aTg7S3TJgvSr8gxGX8,1968
+smftools/informatics/archived/helpers/archived/plot_bed_histograms.py,sha256=78i0mYFuElTPGA2Dt1feO6Z4Grh1Nro3m-F8D5FRBOw,9914
+smftools/informatics/archived/helpers/archived/separate_bam_by_bc.py,sha256=pCLev0OQji1jBdVr25lI_gt9fsozSG8vh7TQkE_UHnY,1800
+smftools/informatics/archived/helpers/archived/split_and_index_BAM.py,sha256=Q7I5qJ5JjW6mSKysfl9NdlFZ6LIy3C8G5rGmG7cn2eA,1224
 smftools/machine_learning/__init__.py,sha256=cWyGN_QVcssqBr_VVr7xh2Inz0P7ylqUmBBcpMgsK0k,257
 smftools/machine_learning/data/__init__.py,sha256=xbfLE-gNjdgxvZ9LKTdvjAtbIHOcs2TR0Gz3YRFbo38,113
 smftools/machine_learning/data/anndata_data_module.py,sha256=ktrdMVMk5yhIUrnu-G_Xf3y7G-KP9PyhYZhobv8TCVg,10063
@@ -105,17 +116,18 @@ smftools/machine_learning/utils/__init__.py,sha256=yOpzBc9AXbarSRfN8Ixh2Z1uWLGpg
 smftools/machine_learning/utils/device.py,sha256=GITrULOty2Fr96Bqt1wi1PaYl_oVgB5Z99Gfn5vQy4o,274
 smftools/machine_learning/utils/grl.py,sha256=BWBDp_kQBigrUzQpRbZzgpfr_WOcd2K2V3MQL-aAIc4,334
 smftools/plotting/__init__.py,sha256=7T3-hZFgTY0nfQgV4J6Vn9ogwkNMlY315kguZR7V1AI,866
-smftools/plotting/autocorrelation_plotting.py,sha256=wdqQ4dawibgZeXRs_G4WterkqOgxHWrJlgZ4PdtO-OA,27456
+smftools/plotting/autocorrelation_plotting.py,sha256=cF9X3CgKiwzL79mgMUFO1tSqdybDoPN1COQQ567InCY,27455
 smftools/plotting/classifiers.py,sha256=8_zabh4NNB1_yVxLD22lfrfl5yfzbEoG3XWqlIqdtrQ,13786
-smftools/plotting/general_plotting.py,sha256=JOeF_lq2hCDt1Vgy8KYLKSzNj8SXGi3z6qMt2P68TDU,31458
+smftools/plotting/general_plotting.py,sha256=2JzE7agm_tILpQ67BHs5pdyPRsHBwcENZe7n4gfMWgM,61350
 smftools/plotting/hmm_plotting.py,sha256=3Eq82gty_0b8GkSMCQgUlbKfzR9h2fJ5rZkB8yYGX-M,10934
 smftools/plotting/position_stats.py,sha256=4XukYIWeWZ_aGSZg1K0t37KA2aknjNNKT5kcKFfuz8Q,17428
 smftools/plotting/qc_plotting.py,sha256=q5Ri0q89udvNUFUNxHzgk9atvQYqUkqkS5-JFq9EqoI,10045
-smftools/preprocessing/__init__.py,sha256=VqhiwJg57m0ePCRAGfX3cJniNLV2jNJpoXZEM2j-0wU,1687
+smftools/preprocessing/__init__.py,sha256=GAQBULUH7fGVabzK5Cq5Wj-0ew0vNA-jWQtR5LAowvs,1746
 smftools/preprocessing/add_read_length_and_mapping_qc.py,sha256=zD_Kxw3DvyOypfuSMGv0ESyt-02w4XlAAMqQxb7yDNQ,5700
-smftools/preprocessing/append_base_context.py,sha256=ohtdHNS1Y9ttLvhLKSwrOyar7HyU2Dw0Ach9WVx5QM8,6221
-smftools/preprocessing/append_binary_layer_by_base_context.py,sha256=I3iiZkVqqB1KqSiA-s-ctl-ESkuTpd7Ot82M0xv_Cm4,6202
-smftools/preprocessing/binarize_on_Youden.py,sha256=O5E3vFc2zXMfKW0p0JGDlmRKEx2_VP6dAqfvrumzz00,1797
+smftools/preprocessing/append_base_context.py,sha256=wGBAADePnys8DLUR15MpRe2BUcfCMDJWaCDDNyjn6AU,6209
+smftools/preprocessing/append_binary_layer_by_base_context.py,sha256=s-7t-VKCs9Y67pX7kH6DNCEkC-RW4nM-UPsBQV2ZwtE,6186
+smftools/preprocessing/binarize.py,sha256=6Vr7Z8zgtJ5rS_uPAx1n3EnQR670V33DlZ_95JmOeWc,484
+smftools/preprocessing/binarize_on_Youden.py,sha256=HGs4p7XiOSYU3_z8QswNHIA9HlrI-7Pp1Kggrn6yUnI,1834
 smftools/preprocessing/binary_layers_to_ohe.py,sha256=Lxd8knelNTaUozfGMFNMlnrOb6uP28Laj3Ymw6cRHL0,1826
 smftools/preprocessing/calculate_complexity.py,sha256=cXMpFrhkwkPipQo2GZGT5yFknMYUMt1t8gz0Cse1DrA,3288
 smftools/preprocessing/calculate_complexity_II.py,sha256=DGfl0jkuBPUpzhKVItN0W7EPzh-QYuR4IxRObPE6gAQ,9301
@@ -123,14 +135,14 @@ smftools/preprocessing/calculate_consensus.py,sha256=6zRpRmb2xdfDu5hctZrReALRb7P
 smftools/preprocessing/calculate_coverage.py,sha256=4WTILzKLzxGLSsQrZkshXP-IRQpoVu3Fkqc0QTpux3Y,2132
 smftools/preprocessing/calculate_pairwise_differences.py,sha256=5zJbNNaFld5qgKRoPyplCmMHflbvAQ9eKWCXPXPpJ60,1774
 smftools/preprocessing/calculate_pairwise_hamming_distances.py,sha256=e5Mzyex7pT29H2PY014uU4Fi_eewbut1JkzC1ffBbCg,961
-smftools/preprocessing/calculate_position_Youden.py,sha256=9GY_WWwaxpB2Xppck3WT1zHtFOhTXrpuDIgbxLC9A7E,7450
+smftools/preprocessing/calculate_position_Youden.py,sha256=yaSd6UDXPCddoN1UR6LgTqE5teJ79Ldw0BAlemc9fB4,7453
 smftools/preprocessing/calculate_read_length_stats.py,sha256=gNNePwMqYZJidzGgT1ZkfSlvc5Y3I3bi5KNYpP6wQQc,4584
-smftools/preprocessing/calculate_read_modification_stats.py,sha256=fQYtwsGt6zq7QBlWtAEaFOkbV_4yXjrj9GnBryEEztc,4779
+smftools/preprocessing/calculate_read_modification_stats.py,sha256=mIlLBqNflVIkuoLxhbyujq3JEKyPl8iebhUlikB9brM,4775
 smftools/preprocessing/clean_NaN.py,sha256=IOcnN5YF05gpPQc3cc3IS83petCnhCpkYiyT6bXEyx0,1937
 smftools/preprocessing/filter_adata_by_nan_proportion.py,sha256=GZcvr2JCsthX8EMw34S9-W3fc6JElw6ka99Jy6f2JvA,1292
 smftools/preprocessing/filter_reads_on_length_quality_mapping.py,sha256=93LgTy_vsPnOZgoiXhZ1-w_pix2oFdBk-dsBUoz33Go,7379
-smftools/preprocessing/filter_reads_on_modification_thresholds.py,sha256=wOmHhQj3xQALQdtQ4-v4POEOat5bEJa-BVmzEE_yrKA,19403
-smftools/preprocessing/flag_duplicate_reads.py,sha256=D7KrDuyy_TSgGvB5aRRmY01k36p92n48YEwmwsUd3IY,65595
+smftools/preprocessing/filter_reads_on_modification_thresholds.py,sha256=4TUvChkSH8R4p_0TpRCh7TounkdUgQHh71TGNmsZ29A,19355
+smftools/preprocessing/flag_duplicate_reads.py,sha256=MySI9En6xVp0FqL7hfiLw0EP3JnGVJWM_yZfkvN-m1U,65585
 smftools/preprocessing/invert_adata.py,sha256=HYMJ1sR3Ui8j6bDjY8OcVQOETzZV-_rrpIYaWLZL6S4,1049
 smftools/preprocessing/load_sample_sheet.py,sha256=AjJf2MrqGHJJ2rNjYi09zV1QkLTq8qGaHGVklXHnPuU,1908
 smftools/preprocessing/make_dirs.py,sha256=lWHXpwC76MFM5sSme9i_WeYUaxutzybendokhny03ds,537
@@ -154,8 +166,8 @@ smftools/tools/archived/classify_methylated_features.py,sha256=Z0N2UKw3luD3CTQ8w
 smftools/tools/archived/classify_non_methylated_features.py,sha256=IJERTozEs7IPL7K-VIjq2q2K36wRCW9iiNSYLAXasrA,3256
 smftools/tools/archived/subset_adata_v1.py,sha256=qyU9iCal03edb5aUS3AZ2U4TlL3uQ42jGI9hX3QF7Fc,1047
 smftools/tools/archived/subset_adata_v2.py,sha256=OKZoUpvdURPtckIQxGTWmOI5jLa-_EU62Xs3LyyehnA,1880
-smftools-0.2.1.dist-info/METADATA,sha256=MXyiJbt1w_Ln4ENxQNbLU0JWwE-S6z1oNZkd8gkf3J8,8958
-smftools-0.2.1.dist-info/WHEEL,sha256=1yFddiXMmvYK7QYTqtRNtX66WJ0Mz8PYEiEUoOUUxRY,87
-smftools-0.2.1.dist-info/entry_points.txt,sha256=NflK6zRv2zlvnjCnDSHycp9w9CczHLfGz9zAc4FtI0I,46
-smftools-0.2.1.dist-info/licenses/LICENSE,sha256=F8LwmL6vMPddaCt1z1S83Kh_OZv50alTlY7BvVx1RXw,1066
-smftools-0.2.1.dist-info/RECORD,,
+smftools-0.2.3.dist-info/METADATA,sha256=w_PRsBPndPoTQZviW9WTuiZV1Pk3ukeJ155OvC4E57M,8787
+smftools-0.2.3.dist-info/WHEEL,sha256=1yFddiXMmvYK7QYTqtRNtX66WJ0Mz8PYEiEUoOUUxRY,87
+smftools-0.2.3.dist-info/entry_points.txt,sha256=q4hg4w-mKkI2leekM_-YZc5XRJzp96Mh1FcU3hac82g,52
+smftools-0.2.3.dist-info/licenses/LICENSE,sha256=F8LwmL6vMPddaCt1z1S83Kh_OZv50alTlY7BvVx1RXw,1066
+smftools-0.2.3.dist-info/RECORD,,

smftools-0.2.3.dist-info/entry_points.txt

@@ -0,0 +1,2 @@
+[console_scripts]
+smftools = smftools.cli_entry:cli

smftools/cli.py

@@ -1,184 +0,0 @@
-import click
-import pandas as pd
-from pathlib import Path
-from typing import Dict, Optional
-
-from . import load_adata
-from .readwrite import merge_barcoded_anndatas_core, safe_read_h5ad, safe_write_h5ad
-
-@click.group()
-def cli():
-    """Command-line interface for smftools."""
-    pass
-
-####### Main processing workflow ###########
-@cli.command()
-@click.argument("config_path", type=click.Path(exists=True))
-def load(config_path):
-    """Load and process data from CONFIG_PATH."""
-    load_adata(config_path)
-##########################################
-
-
-####### Merging existing anndatas from an experiment that used two different demultiplexing rules #######
-REQUIRED_KEYS = ("adata_single_path", "adata_double_path")
-OPTIONAL_KEYS = (
-    "adata_single_backups_path",
-    "adata_double_backups_path",
-    "output_path",
-    "merged_filename",
-)
-
-def _read_config_csv(csv_path: Path) -> Dict[str, str]:
-    """
-    Read a multi-row, two-column CSV of key,value pairs into a dict.
-
-    Supported features:
-      - Optional header ("key,value") or none.
-      - Comments starting with '#' and blank lines are ignored.
-      - If duplicate keys occur, the last one wins.
-      - Keys are matched literally against REQUIRED_KEYS/OPTIONAL_KEYS.
-    """
-    try:
-        # Read as two columns regardless of header; comments ignored.
-        df = pd.read_csv(
-            csv_path,
-            dtype=str,
-            comment="#",
-            header=None,          # treat everything as rows; we'll normalize below
-            usecols=[0, 1],
-            names=["key", "value"]
-        )
-    except Exception as e:
-        raise click.ClickException(f"Failed to read CSV: {e}") from e
-
-    # Drop completely empty rows
-    df = df.fillna("").astype(str)
-    df["key"] = df["key"].str.strip()
-    df["value"] = df["value"].str.strip()
-    df = df[(df["key"] != "") & (df["key"].notna())]
-
-    if df.empty:
-        raise click.ClickException("Config CSV is empty after removing comments/blank lines.")
-
-    # Remove an optional header row if present
-    if df.iloc[0]["key"].lower() in {"key", "keys"}:
-        df = df.iloc[1:]
-        df = df[(df["key"] != "") & (df["key"].notna())]
-        if df.empty:
-            raise click.ClickException("Config CSV contains only a header row.")
-
-    # Build dict; last occurrence of a key wins
-    cfg = {}
-    for k, v in zip(df["key"], df["value"]):
-        cfg[k] = v
-
-    # Validate required keys
-    missing = [k for k in REQUIRED_KEYS if not cfg.get(k)]
-    if missing:
-        raise click.ClickException(
-            "Missing required keys in CSV: "
-            + ", ".join(missing)
-            + "\nExpected keys:\n  - "
-            + "\n  - ".join(REQUIRED_KEYS)
-            + "\nOptional keys:\n  - "
-            + "\n  - ".join(OPTIONAL_KEYS)
-        )
-
-    return cfg
-
-def _resolve_output_path(cfg: Dict[str, str], single_path: Path, double_path: Path) -> Path:
-    """Decide on the output .h5ad path based on CSV; create directories if needed."""
-    merged_filename = cfg.get("merged_filename") or f"merged_{single_path.stem}__{double_path.stem}.h5ad"
-    if not merged_filename.endswith(".h5ad"):
-        merged_filename += ".h5ad"
-
-    output_path_raw = cfg.get("output_path", "").strip()
-
-    if not output_path_raw:
-        out_dir = Path.cwd() / "merged_output"
-        out_dir.mkdir(parents=True, exist_ok=True)
-        return out_dir / merged_filename
-
-    output_path = Path(output_path_raw)
-
-    if output_path.suffix.lower() == ".h5ad":
-        output_path.parent.mkdir(parents=True, exist_ok=True)
-        return output_path
-
-    # Treat as directory
-    output_path.mkdir(parents=True, exist_ok=True)
-    return output_path / merged_filename
-
-def _maybe_read_adata(label: str, primary: Path, backups: Optional[Path]):
-
-    if backups:
-        click.echo(f"Loading {label} from {primary} with backups at {backups} ...")
-        return safe_read_h5ad(primary, backups_path=backups, restore_backups=True)
-    else:
-        click.echo(f"Loading {label} from {primary} with backups disabled ...")
-        return safe_read_h5ad(primary, restore_backups=False)
-
-
-@cli.command()
-@click.argument("config_path", type=click.Path(exists=True, dir_okay=False, readable=True, path_type=Path))
-def merge_barcoded_anndatas(config_path: Path):
-    """
-    Merge two AnnData objects from the same experiment that were demultiplexed
-    under different end-barcoding requirements, using a 1-row CSV for config.
-
-    CSV must include:
-      - adata_single_path
-      - adata_double_path
-
-    Optional columns:
-      - adata_single_backups_path
-      - adata_double_backups_path
-      - output_path            (file or directory; default: ./merged_output/)
-      - merged_filename        (default: merged_<single>__<double>.h5ad)
-
-    Example CSV:
-
-        adata_single_path,adata_double_path,adata_single_backups_path,adata_double_backups_path,output_path,merged_filename
-        /path/single.h5ad,/path/double.h5ad,,,,merged_output,merged_run.h5ad
-    """
-    try:
-        cfg = _read_config_csv(config_path)
-
-        single_path = Path(cfg["adata_single_path"]).expanduser().resolve()
-        double_path = Path(cfg["adata_double_path"]).expanduser().resolve()
-
-        for p, label in [(single_path, "adata_single_path"), (double_path, "adata_double_path")]:
-            if not p.exists():
-                raise click.ClickException(f"{label} does not exist: {p}")
-
-        single_backups = Path(cfg["adata_single_backups_path"]).expanduser().resolve() if cfg.get("adata_single_backups_path") else None
-        double_backups = Path(cfg["adata_double_backups_path"]).expanduser().resolve() if cfg.get("adata_double_backups_path") else None
-
-        if single_backups and not single_backups.exists():
-            raise click.ClickException(f"adata_single_backups_path does not exist: {single_backups}")
-        if double_backups and not double_backups.exists():
-            raise click.ClickException(f"adata_double_backups_path does not exist: {double_backups}")
-
-        output_path = _resolve_output_path(cfg, single_path, double_path)
-
-        # Load
-        adata_single, read_report_single = _maybe_read_adata("single-barcoded AnnData", single_path, single_backups)
-        adata_double, read_report_double = _maybe_read_adata("double-barcoded AnnData", double_path, double_backups)
-
-        click.echo("Merging AnnDatas ...")
-        merged = merge_barcoded_anndatas_core(adata_single, adata_double)
-
-        click.echo(f"Writing merged AnnData to: {output_path}")
-        backup_dir = output_path.cwd() / "merged_backups"
-        safe_write_h5ad(merged, output_path, backup=True, backup_dir=backup_dir)
-
-        click.secho(f"Done. Merged AnnData saved to {output_path}", fg="green")
-
-    except click.ClickException:
-        raise
-    except Exception as e:
-        # Surface unexpected errors cleanly
-        raise click.ClickException(f"Unexpected error: {e}") from e
-
-################################################################################################################

smftools/informatics/fast5_to_pod5.py

@@ -1,24 +0,0 @@
-# fast5_to_pod5
-
-def fast5_to_pod5(fast5_dir, output_pod5='FAST5s_to_POD5.pod5'):
-    """
-    Convert Nanopore FAST5 files to POD5 file
-
-    Parameters:
-        fast5_dir (str): String representing the file path to a directory containing all FAST5 files to convert into a single POD5 output.
-        output_pod5 (str): The name of the output POD5.
-    
-    Returns:
-        None
-    
-    """
-    import subprocess
-    from pathlib import Path
-
-    if isinstance(fast5_dir, (list, tuple)):
-        cmd = ["pod5", "convert", "fast5"] + fast5_dir + ["--output", output_pod5]
-        subprocess.run(cmd)
-    elif Path(fast5_dir).is_file():
-        subprocess.run(["pod5", "convert", "fast5", fast5_dir, "--output", output_pod5])
-    elif Path(fast5_dir).is_dir():
-        subprocess.run(["pod5", "convert", "fast5", f".{fast5_dir}*.fast5", "--output", output_pod5])

smftools/informatics/helpers/__init__.py

@@ -1,73 +0,0 @@
-from .align_and_sort_BAM import align_and_sort_BAM
-from .aligned_BAM_to_bed import aligned_BAM_to_bed
-from .bam_qc import bam_qc
-from .bed_to_bigwig import bed_to_bigwig
-from .binarize_converted_base_identities import binarize_converted_base_identities
-from .canoncall import canoncall
-from .complement_base_list import complement_base_list
-from .converted_BAM_to_adata_II import converted_BAM_to_adata_II
-from .concatenate_fastqs_to_bam import concatenate_fastqs_to_bam
-from .count_aligned_reads import count_aligned_reads
-from .demux_and_index_BAM import demux_and_index_BAM
-from .discover_input_files import *
-from .extract_base_identities import extract_base_identities
-from .extract_mods import extract_mods
-from .extract_read_features_from_bam import extract_read_features_from_bam
-from .extract_read_lengths_from_bed import extract_read_lengths_from_bed
-from .extract_readnames_from_BAM import extract_readnames_from_BAM
-from .find_conversion_sites import find_conversion_sites
-from .generate_converted_FASTA import convert_FASTA_record, generate_converted_FASTA
-from .get_chromosome_lengths import get_chromosome_lengths
-from .get_native_references import get_native_references
-from .index_fasta import index_fasta
-from .make_dirs import make_dirs
-from .make_modbed import make_modbed
-from .modcall import modcall
-from .modkit_extract_to_adata import modkit_extract_to_adata
-from .modQC import modQC
-from .one_hot_encode import one_hot_encode
-from .ohe_batching import ohe_batching
-from .one_hot_decode import one_hot_decode
-from .ohe_layers_decode import ohe_layers_decode
-from .plot_bed_histograms import plot_bed_histograms
-from .run_multiqc import run_multiqc
-from .separate_bam_by_bc import separate_bam_by_bc
-from .split_and_index_BAM import split_and_index_BAM
-
-__all__ = [
-    "align_and_sort_BAM",
-    "aligned_BAM_to_bed",
-    "bam_qc",
-    "bed_to_bigwig",
-    "binarize_converted_base_identities",
-    "canoncall",
-    "complement_base_list",
-    "converted_BAM_to_adata_II",
-    "concatenate_fastqs_to_bam",
-    "count_aligned_reads",
-    "demux_and_index_BAM",
-    "extract_base_identities",
-    "extract_mods",
-    "extract_read_features_from_bam",
-    "extract_read_lengths_from_bed",
-    "extract_readnames_from_BAM",
-    "find_conversion_sites",
-    "convert_FASTA_record",
-    "generate_converted_FASTA",
-    "get_chromosome_lengths",
-    "get_native_references",
-    "index_fasta",
-    "make_dirs",
-    "make_modbed",
-    "modcall",
-    "modkit_extract_to_adata",
-    "modQC",
-    "one_hot_encode",
-    "ohe_batching",
-    "one_hot_decode",
-    "ohe_layers_decode",
-    "plot_bed_histograms",
-    "run_multiqc",
-    "separate_bam_by_bc",
-    "split_and_index_BAM"
-]

smftools/informatics/helpers/align_and_sort_BAM.py

@@ -1,86 +0,0 @@
-## align_and_sort_BAM
-
-def align_and_sort_BAM(fasta, 
-                       input, 
-                       bam_suffix='.bam', 
-                       output_directory='aligned_outputs', 
-                       make_bigwigs=False, 
-                       threads=None, 
-                       aligner='minimap2',
-                       aligner_args=['-a', '-x', 'map-ont', '--MD', '-Y', '-y', '-N', '5', '--secondary=no']):
-    """
-    A wrapper for running dorado aligner and samtools functions
-    
-    Parameters:
-        fasta (str): File path to the reference genome to align to.
-        input (str): File path to the basecalled file to align. Works for .bam and .fastq files
-        bam_suffix (str): The suffix to use for the BAM file.
-        output_directory (str): A file path to the directory to output all the analyses.
-        make_bigwigs (bool): Whether to make bigwigs
-        threads (int): Number of additional threads to use
-        aligner (str): Aligner to use. minimap2 and dorado options
-        aligner_args (list): list of optional parameters to use for the alignment
-
-    Returns:
-        None
-            The function writes out files for: 1) An aligned BAM, 2) and aligned_sorted BAM, 3) an index file for the aligned_sorted BAM, 4) A bed file for the aligned_sorted BAM, 5) A text file containing read names in the aligned_sorted BAM
-    """
-    import subprocess
-    import os
-
-    input_basename = os.path.basename(input)
-    input_suffix = '.' + input_basename.split('.')[1]
-    input_as_fastq = input_basename.split('.')[0] + '.fastq'
-
-    output_path_minus_suffix = os.path.join(output_directory, input_basename.split(input_suffix)[0])
-    
-    aligned_BAM=f"{output_path_minus_suffix}_aligned"
-    aligned_sorted_BAM=f"{aligned_BAM}_sorted"
-    aligned_output = aligned_BAM + bam_suffix
-    aligned_sorted_output = aligned_sorted_BAM + bam_suffix
-
-    if threads:
-        threads = str(threads)
-    else:
-        pass
-    
-    if aligner == 'minimap2':
-        print(f"Converting BAM to FASTQ: {input}")
-        bam_to_fastq_command = ['samtools', 'fastq', input]
-        subprocess.run(bam_to_fastq_command, stdout=open(input_as_fastq, "w"))
-        print(f"Aligning FASTQ to Reference: {input_as_fastq}")
-        if threads:
-            minimap_command = ['minimap2'] + aligner_args + ['-t', threads, fasta, input_as_fastq]
-        else:
-            minimap_command = ['minimap2'] + aligner_args + [fasta, input_as_fastq]
-        subprocess.run(minimap_command, stdout=open(aligned_output, "w"))
-        os.remove(input_as_fastq)
-
-    elif aligner == 'dorado':
-        # Run dorado aligner
-        print(f"Aligning BAM to Reference: {input}")
-        if threads:
-            alignment_command = ["dorado", "aligner", "-t", threads] + aligner_args + [fasta, input]
-        else:
-            alignment_command = ["dorado", "aligner"] + aligner_args + [fasta, input]
-        subprocess.run(alignment_command, stdout=open(aligned_output, "w"))
-
-    else:
-        print(f'Aligner not recognized: {aligner}. Choose from minimap2 and dorado')
-        return
-
-    # Sort the BAM on positional coordinates
-    print(f"Sorting BAM: {aligned_output}")
-    if threads:
-        sort_command = ["samtools", "sort", "-@", threads, "-o", aligned_sorted_output, aligned_output]
-    else:
-        sort_command = ["samtools", "sort", "-o", aligned_sorted_output, aligned_output]
-    subprocess.run(sort_command)
-
-    # Create a BAM index file
-    print(f"Indexing BAM: {aligned_sorted_output}")
-    if threads:
-        index_command = ["samtools", "index", "-@", threads, aligned_sorted_output]
-    else:
-        index_command = ["samtools", "index", aligned_sorted_output]
-    subprocess.run(index_command)