From 2d8e72054d709427fd16f11e267b1207502f61aa Mon Sep 17 00:00:00 2001 From: "Dr. Kashif Rasul" Date: Tue, 19 May 2020 09:49:51 +0200 Subject: [PATCH] Revert "Enabling loading one line at a time" This reverts commit ba95f937560aaeffa45b90189653d81200ab4d14. --- pts/dataset/file_dataset.py | 26 ++++---------------------- 1 file changed, 4 insertions(+), 22 deletions(-) diff --git a/pts/dataset/file_dataset.py b/pts/dataset/file_dataset.py index cd4ae52..e48d8bf 100644 --- a/pts/dataset/file_dataset.py +++ b/pts/dataset/file_dataset.py @@ -53,19 +53,13 @@ class JsonLinesFile: def __init__(self, path: Path, shuffle: bool = True) -> None: self.path = path - if shuffle: - self.__iter_internal = self.__iter_shuffle__ - else: - self.__iter_internal = self.__iter_inorder__ - - def __iter__(self): - return self.__iter_internal() + self.shuffle = shuffle - def __iter_shuffle__(self): + def __iter__(self): with open(self.path) as jsonl_file: lines = jsonl_file.read().splitlines() - - random.shuffle(lines) + if self.shuffle: + random.shuffle(lines) for line_number, raw in enumerate(lines, start=1): span = Span(path=self.path, line=line_number) @@ -73,18 +67,6 @@ class JsonLinesFile: yield Line(json.loads(raw), span=span) except ValueError: raise Exception(f"Could not read json line {line_number}, {raw}") - - def __iter_inorder__(self): - with open(self.path) as jsonl_file: - line_number = 0 - while True: - line_number += 1 - raw = jsonl_file.readline() - span = Span(path=self.path, line=line_number) - try: - yield Line(json.loads(raw), span=span) - except ValueError: - raise Exception(f"Could not read json line {line_number}, {raw}") def __len__(self): # 1MB