标签中提取纯文本rD r8 rY br
r< r= r[ r9 r\ r] z) r7 r3 TrK [)spanstrongbemufontz[ \t]+ )r_ children
isinstancer r? r, rM r rA r r` r' subrc ) p_tagrE rw childtr= r[ rx a_hrefa_text full_hrefs r rf rf s zz7)
C 9 e_ -E
"Aq3w
ZZ4
4KC
ZZ5
))E2&C))E2&C"%.."83ghPS>TC58*B//
ZZ3
YYvr*F^^$^/F&&,&7&7&?FWXW]E^ 6("YKr22
ZZE
E 5>>#T>2S8 8C)9*
&&C
%
+
+
-C
++fc
"C
&&C
%
+
+
-C Jr c J t j d } | j dt dd | j }|j
dkD r |j
nd}g }t
d|dz D ] }|dk( rt d t d
}nt d t d | d }t d
| d | d t | }t | }|s
t d n@t dt | d |j | t j d t dt | d |s t d y g } t# | D ] \ }
} t d|
dz d t | d| d d d d t% | \ } }
}}| s| d } t' j( dd| j+ } |r|d d nd }| d | d | |||
t, |rt/ j0 |d nd d!}| j3 | t j d" t d#t | d$ | rt5 | t, t7 }|rw |j9 d%t, f }|j; d }|j9 d&t, f }|j; d }t d'| d(| d) |j= t d+ y # t $ r}t d| Y d }~ d }~ww xY w# t $ r}t d*| Y d }~^d }~ww xY w# |j= w xY w),Nr ) descriptionz--pagesr u 页数，0=全量53页)typedefaulthelp5 r" /z/xxgklist.shtmlz
/xxgklist_z.shtmlzPage z...z
empty, stopz found z itemsz WARN: g333333?z
Total list: zNo items found, abortz [z] r% ( z\s+rY i r8 r F) ensure_ascii)r
source_urlr% rP summaryr& site_namerq g?z
Inserting z records...z.SELECT COUNT(*) FROM gov_raw WHERE site_name=?z1SELECT COUNT(*) FROM gov_search WHERE site_name=?zDB: z rows (FTS: z rows)z WARN DB query: DONE) argparseArgumentParser add_argumentint
parse_argspagesranger- LIST_PATHr r r4 lenextendr timesleep enumeratery r' r r, SITE_NAMEjsondumpsr. r r r fetchoneclose)parserargs max_pages all_itemspager r/ r0 r recordsirj r% r& rP rq r recordr cursor raw_count fts_counts r mainr s
$
$1Y
ZF
Q=UV D "jj1n
"IIaQ ' 19 Ja {/:C Ja {*TF&AC
dV1YKs+,
c?D t$E
' HSZL/ 0 U #
3%(
N3y>*&
12
%&G Y' 4
AaC5#i.)DM#2,>+?sCD0