+ 4jRt^RIHt^RIt^RIt^RIt^RIHtH t ^RI t ^RI t ^RI H t HtHtHt^RIHtHtHtHtHtHt]'d^RIHtR.t!RR 4t!R R4tR#) z/httpx-based async HTTP transport for nab-index.) annotationsN) TYPE_CHECKINGAny) MAX_REDIRECTS MAX_RETRIESRETRY_STATUSES next_delay)DEFAULT_HEADERSContentDecodingError HttpError accepts_gzip decode_bodyraise_for_error_status)MappingHttpxAsyncTransportc]tRt^tRtRtRRlt]RRl4t]RRl4t ]RR l4t ]R R l4t ]R R l4t RRlt RRltRtR#)_HttpxResponseafAdapter that gives an httpx response the HttpResponse shape. The body is fetched undecoded, and decoded by the transport when the request asked for gzip (see :func:`~nab_index.transport.decode_body`), so it is carried here rather than read from the httpx response. ``raise_for_status`` follows the shared 4xx/5xx rule; httpx's own raises on every non-2xx. c$V^8dQhRRRRRR/#)responsezhttpx.ResponsecontentbytesreturnNone)formats"q/home/user/pdf2html-billing-ledger-validate/.venv/lib/python3.14/site-packages/nab_index/httpx_async_transport.py __annotate___HttpxResponse.__annotate__,s!   % D c WnW nR#N) _response_content)selfrrs&&&r__init___HttpxResponse.__init__,s ! rcV^8dQhRR/#)rrintr)rs"rrr1s**S*rc .VPP#r!)r" status_coder$s&rr*_HttpxResponse.status_code0s~~)))rcV^8dQhRR/#rrstrr)rs"rrr5s''S'rc @\VPP4#r!)r/r"urlr+s&rr1_HttpxResponse.url4s4>>%%&&rcV^8dQhRR/#)rrzMapping[str, str]r)rs"rrr9s&&*&rc .VPP#r!)r"headersr+s&rr5_HttpxResponse.headers8s~~%%%rcV^8dQhRR/#)rrrr)rs"rrr=src VP#r!)r#r+s&rr_HttpxResponse.content<s }}rcV^8dQhRR/#r.r)rs"rrrAs--c-rc 8VPPR4#)zutf-8)r#decoder+s&rtext_HttpxResponse.text@s}}##G,,rcV^8dQhRR/#)rrrr)rs"rrrDs**c*rc B\P!VP4#r!)_jsonloadsr#r+s&rjson_HttpxResponse.jsonDs{{4==))rcV^8dQhRR/#rrrr)rs"rrrGsEE$Erc Z\VPPVP4R#r!)rr"r*r1r+s&rraise_for_status_HttpxResponse.raise_for_statusGst~~99488Dr)r#r"N)__name__ __module__ __qualname____firstlineno____doc__ __slots__r%propertyr*r1r5rr=rCrH__static_attributes__rrrrrs*I **''&&--*EErrcN]tRt^KtRtRR/RRlltRR/RR lltR R ltR tR#) rz]Async HTTP transport using httpx. HTTP/2 is enabled by default for connection multiplexing. http2Tc V^8dQhRRRR/#)rrSboolrrr)rs"rr HttpxAsyncTransport.__annotate__Qs       rc \P!VR\\P!\ P 4R7VnR#)zCreate a transport.T)rSfollow_redirects max_redirectsverifyN)httpx AsyncClientr truststore SSLContextsslPROTOCOL_TLS_CLIENT_client)r$rSs&$rr%HttpxAsyncTransport.__init__Qs3(("'(()@)@A   rr5Nc$V^8dQhRRRRRR/#)rr1r/r5zdict[str, str] | Nonerrr)rs"rrrV\s$:':':'$9:' :'rc ~"\\4pVeVPV4\V4p^pVPP RWR7;_uu_4GRjxL pRP VP4Uu.uRjxL qwNK L/L D Muupiup4pRRR4GRjxL M +GRjxL 'giM;iT'd&\XXPPR44MXp XP\9d \Yi4#T^, pT\8d \Yi4#\YVPPR44p M \ P"\ P$3dp RT RT 2p \'T 4T hRp ? i\ P(\*3d>p T^, pT\8dRT RT 2p \'T 4T h\T4p Rp ? M*Rp ? i\,dp RT RT 2p \'T 4T hRp ? ii;i\.P0!T 4GRjxL EK 5i) aSend a GET request, retrying a blip. httpx has no retry machinery beyond reconnecting, so the shared policy runs in this loop rather than in the client. The body is read raw and decoded with ``decode_body``, so a truncated gzip stream is retried instead of returned as if complete. ``headers`` overrides entries of :data:`~nab_index.transport.DEFAULT_HEADERS`, so a caller can pass :data:`~nab_index.transport.IDENTITY_HEADERS` to get the body undecoded. NGET)r5rzContent-Encodingz Retry-AfterzGET z failed: )dictr updater rastreamjoin aiter_rawr r5getr*rrrrr[TooManyRedirectsUnsupportedProtocolr HTTPErrorr Exceptionasynciosleep) r$r1r5request_headersr<failuresrpartrawrdelayexcmsgs &&$ rrkHttpxAsyncTransport.get\s/    " "7 +o.# R<<..3/((8;M;M;O#P#P4D$Q#PQC  X%5%5%9%9:L%MN.''~=)(<<A k))(<<"8-=-=-A-A--PQ3**E,E,EF .SE3%0n#-OO%9: -A k) Yse4C#C.c1"8, .SE3%0n#-  . --& & &s3H=)E"B E"#B3B BBB B E"BB  B3! E",B/-E"3C 9B<: C C E"'E">A$H="%HFH8H93G1,H=1 H>H?HHH=4H75H=cV^8dQhRR/#rFr)rs"rrrVs$$d$rc V"VPP4GRjxL R#L5i)zClose the underlying client.N)raacloser+s&rr|HttpxAsyncTransport.aclosesll!!###s )'))ra) rJrKrLrMrNr%rkr|rQrrrrrKs+    :'<@:'x$$r)rN __future__rrprCrAr_typingrrr[r]retryrrrr transportr r r r r rcollections.abcr__all__rrrrrrsZ5" % II'  )E)EXO$O$r