simon987/Simple-Incremental-Search-Tool
1
0
Fork 0
mirror of https://github.com/simon987/Simple-Incremental-Search-Tool.git synced 2025-12-16 00:29:05 +00:00
Code Issues Projects Releases Wiki Activity

Audio tags in search page and svg thumbnail generation

Browse Source
This commit is contained in:
simon
2018-04-17 11:45:31 -04:00
parent dff7ddc511
commit 4eb9cf6b63
8 changed files with 128 additions and 105 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
crawler.py
View File

@@ -5,7 +5,7 @@ from multiprocessing import Process, Value
from apscheduler.schedulers.background import BackgroundScheduler
from parsing import GenericFileParser, Md5CheckSumCalculator, ExtensionMimeGuesser, MediaFileParser, TextFileParser, \
PictureFileParser, Sha1CheckSumCalculator, Sha256CheckSumCalculator, ContentMimeGuesser, MimeGuesser, FontParser, \
PdfFileParser, DocxParser
PdfFileParser, DocxParser, EbookParser
from indexer import Indexer
from search import Search
from thumbnail import ThumbnailGenerator
@@ -77,7 +77,7 @@ class Crawler:
except FileNotFoundError:
continue # File was deleted
if self.indexer is not None:
if self.indexer is not None and len(self.documents) > 0:
self.indexer.index(self.documents, self.dir_id)
def countFiles(self, root_dir: str):
@@ -141,7 +141,8 @@ class TaskManager:
PictureFileParser(chksum_calcs),
FontParser(chksum_calcs),
PdfFileParser(chksum_calcs, int(directory.get_option("TextFileContentLength"))), # todo get content len from other opt
DocxParser(chksum_calcs, int(directory.get_option("TextFileContentLength")))], # todo get content len from other opt
DocxParser(chksum_calcs, int(directory.get_option("TextFileContentLength"))), # todo get content len from other opt
EbookParser(chksum_calcs, int(directory.get_option("TextFileContentLength")))], # todo get content len from other opt
mime_guesser, self.indexer, directory.id)
c.crawl(directory.path, counter)