tommy/libpostal
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

[fix] check the first phrase for components and bail if it matches something other than the specified tag

Browse Source
This commit is contained in:
Al
2016-05-05 12:46:01 -04:00
parent cf0dcd849c
commit f59ee4a136
1 changed files with 11 additions and 8 deletions
Download Patch File Download Diff File Expand all files Collapse all files

19
scripts/geodata/addresses/components.py
View File

@@ -229,24 +229,27 @@ class AddressExpander(object):
join_phrase = six.u(' ') if whitespace else six.u('') join_phrase = six.u(' ') if whitespace else six.u('')
if num_phrases > 0: if num_phrases > 0:
current_phrase_tokens = tokens_lower[current_phrase_start:current_phrase_start + current_phrase_len]
current_phrase = join_phrase.join([t for t, c in current_phrase_tokens])
# Handles cases like addr:city="Harlem" when Harlem is a neighborhood
tags = components.get(current_phrase, set())
if tags and tag not in tags:
return None
# Return phrase with original capitalization # Return phrase with original capitalization
return join_phrase.join([t for t, c in tokens[:total_tokens]]) return join_phrase.join([t for t, c in tokens[:total_tokens]])
elif num_phrases == 0 and total_tokens > 0: elif num_phrases == 0 and total_tokens > 0:
phrase = join_phrase.join([t for t, c in phrase_tokens]) phrase = join_phrase.join([t for t, c in phrase_tokens])
if tag not in components.get(phrase, set()): if tag not in components.get(phrase, set()):
return None return None
elif num_phrases == 0:
current_phrase_tokens = tokens_lower[current_phrase_start:current_phrase_start + current_phrase_len]
current_phrase = join_phrase.join([t for t, c in current_phrase_tokens])
current_phrase_start = total_tokens current_phrase_start = total_tokens
current_phrase_len = len(phrase_tokens) current_phrase_len = len(phrase_tokens)
current_phrase_tokens = tokens_lower[current_phrase_start:current_phrase_start + current_phrase_len]
current_phrase = join_phrase.join([t for t, c in current_phrase_tokens])
# Handles cases like addr:city="Harlem" when Harlem is a neighborhood
tags = components.get(current_phrase, set())
if tags and tag not in tags:
return None
total_tokens += len(phrase_tokens) total_tokens += len(phrase_tokens)
num_phrases += 1 num_phrases += 1
else: else: