diff options
Diffstat (limited to 'scripts/genotypes')
| -rw-r--r-- | scripts/genotypes/preprocess_csv_collect_info.py | 15 |
1 files changed, 6 insertions, 9 deletions
diff --git a/scripts/genotypes/preprocess_csv_collect_info.py b/scripts/genotypes/preprocess_csv_collect_info.py index 771c11d..eaad1d7 100644 --- a/scripts/genotypes/preprocess_csv_collect_info.py +++ b/scripts/genotypes/preprocess_csv_collect_info.py @@ -51,7 +51,7 @@ def identify_columns( return FieldsIdentity(**_id) -def process_transposed(line: tuple[str, ...]) -> LineDetails: +def process_transposed(line: tuple[str, ...], na_strings: tuple[str, ...]) -> LineDetails: """Process a line in a transposed file.""" raise NotImplementedError("Please implement this!") @@ -96,7 +96,7 @@ def file_statistics(lines: Iterator[str], fileconfigs: dict) -> FileStats: if _line.startswith("@"): # custom GeneNetwork Settings stats["dot_geno_lines"] = stats.get("dot_geno_lines", 0) + 1 _fld = tuple(item.strip() for item in line.strip("@").split(":")) - dot_geno[_fld[0]] = dot_geno[_fld[1]] + dotgeno[_fld[0]] = _fld[1] continue stats["data_lines"] = stats.get("data_lines", 0) + 1 @@ -123,14 +123,12 @@ def file_statistics(lines: Iterator[str], fileconfigs: dict) -> FileStats: line_dets = ( process_transposed( fields, - na_strings=fileconfigs["transposed"], - ...) + na_strings=fileconfigs["na_strings"]) if fileconfigs["transposed"] else process_untransposed( fields, - na_strings=fileconfigs["transposed"], - headers, - ...)) + fileconfigs, + headers=headers)) if fileconfigs["transposed"]: headers = headers + (line_dets.header) @@ -139,7 +137,7 @@ def file_statistics(lines: Iterator[str], fileconfigs: dict) -> FileStats: # TODO: do more processing continue # end of line processing - col_ids = identify_columns(fields) + col_ids = identify_columns(headers, fileconfigs) return FileStats( **stats, **({"dot_geno_fields": DotGenoFields(**dotgeno)} @@ -147,7 +145,6 @@ def file_statistics(lines: Iterator[str], fileconfigs: dict) -> FileStats: else {}), samples_list=col_ids.samples, geno_symbols=tuple(genosymbols)) - if __name__ == "__main__": |