X-Git-Url: https://code.communitydata.science/mediawiki_dump_tools.git/blobdiff_plain/9a0c157ebbe6bcfdb7de17ba7938f3fb26d7f32f..b1bea09ad63a3b651e25b0bc36b31e9b7d6a8553:/wikiq?ds=inline diff --git a/wikiq b/wikiq index 8c51979..1a951e8 100755 --- a/wikiq +++ b/wikiq @@ -8,7 +8,7 @@ import argparse import sys import os, os.path import re -from datetime import datetime +from datetime import datetime,timezone from subprocess import Popen, PIPE from collections import deque @@ -141,6 +141,15 @@ class RegexPair(object): if self.has_groups: self.capture_groups = list(self.pattern.groupindex.keys()) + def get_pyarrow_fields(self): + if self.has_groups: + fields = [pa.field(self._make_key(cap_group),pa.list_(pa.string())) + for cap_group in self.capture_groups] + else: + fields = [pa.field(self.label, pa.list_(pa.string()))] + + return fields + def _make_key(self, cap_group): return ("{}_{}".format(self.label, cap_group)) @@ -209,6 +218,7 @@ class RevDataBase(): editor: str = None anon: bool = None + urlencode = False pa_schema_fields = [ pa.field("revid", pa.int64()), pa.field("date_time",pa.timestamp('ms')), @@ -232,7 +242,7 @@ class RevDataBase(): def to_tsv_row(self): row = [] - for f in self.fields(): + for f in dc.fields(self): val = getattr(self, f.name) if getattr(self, f.name) is None: row.append("") @@ -244,36 +254,26 @@ class RevDataBase(): elif f.name in {'editor','title'}: s = '"' + val + '"' - if f.name in TO_ENCODE: - row.append(quote(str(val))) + if self.urlencode and f.name in TO_ENCODE: + row.append(quote(str(s))) + else: + row.append(s) elif f.type == list[int]: row.append('"' + ",".join([str(x) for x in val]) + '"') elif f.type == str: - if f.name in TO_ENCODE: + if self.urlencode and f.name in TO_ENCODE: row.append(quote(str(val))) + else: + row.append(val) else: row.append(val) - return '\t'.join(row) - - # def __init__(revid: int, - # date_time: datetime, - # articleid: int, - # editorid: int, - # title: str, - # namespace: int, - # deleted: bool, - # test_chars: int, - # revert: bool, - # reverteds: list[bool], - # sha1: str, - # minor: bool, - # editor: str, - # anon: bool): + return '\t'.join(map(str,row)) - + def header_row(self): + return '\t'.join(map(lambda f: f.name, dc.fields(self))) @dataclass() class RevDataCollapse(RevDataBase): @@ -341,9 +341,7 @@ class WikiqParser(): self.revdata_type.pa_schema_fields = revdata_type.pa_schema_fields + self.regex_schemas - - # print(list(map(lambda d: d.name, dc.fields(self.revdata_type)))) - # print(self.revdata_type.pa_schema_fields) + self.revdata_type.urlencode = self.urlencode if output_parquet is True: self.output_parquet = True @@ -352,17 +350,22 @@ class WikiqParser(): self.parquet_buffer = [] self.parquet_buffer_size = parquet_buffer_size else: - self.output_file = open(output_file,'w') - + self.print_header = True + if output_file == sys.stdout: + + self.output_file = output_file + else: + self.output_file = open(output_file,'w') + self.output_parquet = False def make_matchmake_pairs(self, patterns, labels): if (patterns is not None and labels is not None) and \ (len(patterns) == len(labels)): result = [] for pattern, label in zip(patterns, labels): - result.append(RegexPair(pattern, label)) - self.regex_schemas.append(pa.field(label, pa.list_(pa.string()))) - + rp = RegexPair(pattern, label) + result.append(rp) + self.regex_schemas = self.regex_schemas + rp.get_pyarrow_fields() return result elif (patterns is None and labels is None): return [] @@ -431,7 +434,7 @@ class WikiqParser(): if self.persist != PersistMethod.none: window = deque(maxlen=PERSISTENCE_RADIUS) - + if self.persist == PersistMethod.sequence: state = mwpersistence.DiffState(SequenceMatcher(tokenizer = wikitext_split), revert_radius=PERSISTENCE_RADIUS) @@ -449,7 +452,7 @@ class WikiqParser(): for rev in page: rev_data = self.revdata_type(revid = rev.id, - date_time = datetime.fromtimestamp(rev.timestamp.unix()), + date_time = datetime.fromtimestamp(rev.timestamp.unix(), tz=timezone.utc), articleid = page.id, editorid = "" if rev.deleted.user == True or rev.user.id is None else rev.user.id, title = page.title, @@ -468,7 +471,6 @@ class WikiqParser(): if rev.sha1: text_sha1 = rev.sha1 else: - text_sha1 = sha1(bytes(rev.text, "utf8")).hexdigest() rev_data.sha1 = text_sha1 @@ -491,8 +493,8 @@ class WikiqParser(): if not rev.deleted.user: # wrap user-defined editors in quotes for fread rev_data.editor = rev.user.text - rev_data.anon = rev.user.id == None - + rev_data.anon = rev.user.id is None + #if re.match(r'^#redirect \[\[.*\]\]', rev.text, re.I): # redirect = True #else: @@ -504,8 +506,8 @@ class WikiqParser(): if self.collapse_user: rev_data.collapsed_revs = rev.collapsed_revs + # get the if self.persist != PersistMethod.none: - if not rev.deleted.text: if self.persist != PersistMethod.legacy: @@ -521,12 +523,12 @@ class WikiqParser(): num_token_revs, num_tokens = calculate_persistence(old_tokens_added) - rev_data.token_revs = num_token_revs - rev_data.tokens_added = num_tokens - rev_data.tokens_removed = len(old_tokens_removed) - rev_data.tokens_window = PERSISTENCE_RADIUS-1 + old_rev_data.token_revs = num_token_revs + old_rev_data.tokens_added = num_tokens + old_rev_data.tokens_removed = len(old_tokens_removed) + old_rev_data.tokens_window = PERSISTENCE_RADIUS-1 - self.print_rev_data(rev_data) + self.print_rev_data(old_rev_data) else: self.print_rev_data(rev_data) @@ -559,7 +561,7 @@ class WikiqParser(): self.pq_writer.close() else: - output_file.close() + self.output_file.close() def write_parquet_row(self, rev_data): @@ -604,7 +606,11 @@ class WikiqParser(): printfunc(rev_data) def write_tsv_row(self, rev_data): - line = rev_data.to_tsv_line() + if self.print_header: + print(rev_data.header_row(), file=self.output_file) + self.print_header = False + + line = rev_data.to_tsv_row() print(line, file=self.output_file) @@ -737,7 +743,6 @@ if len(args.dumpfiles) > 0: regex_comment_label = args.regex_comment_label, output_parquet=output_parquet) - print(wikiq.output_parquet) wikiq.process() # close things