gjj UdZddlZddlZddlmZmZddlmZmZm Z ddl m Z ddl m Z mZddlmZd e jDZeeefed <ejd Zee jZeeed <ehd Zeeed<eddhZeeed<edGddZedGddZdede fdZ!ej"d.dededzdefdZ#dedzde$edzeffdZ%dedzdeefdZ&d.dededzde fdZ'd edefd!Z(d.d"ededzdefd#Z)d$edefd%Z*d&ed'ede$eedze dzffd(Z+d)ed'ede$e j,effd*Z-d)ed+e j,d'edefd,Z.d)ed+e j,d'edefd-Z/dS)/aCentralized parser for Hugging Face Hub URIs ('hf://...') and mount specifications. A HF URI is a URI-like string that identifies a location on the Hugging Face Hub: a model/dataset/space/kernel repository, a bucket, optionally a revision, and optionally a path inside the repo or bucket. Canonical syntax: ``` hf://[/][@][/] ``` For convenience, [`parse_hf_uri`] also accepts Hugging Face **web URLs** (the ones you copy-paste from your browser), e.g. 'https://huggingface.co/datasets/my-org/my-dataset/blob/main/train.csv'. They are normalized to the canonical 'hf://' form before parsing. Only unambiguous URLs (repository / bucket pages and file/folder viewer routes) are accepted; any other route is rejected rather than guessed. A HF mount wraps a HF URI with a local mount path and an optional ':ro'/':rw' flag (used by Spaces and Jobs volumes): ``` hf://[/][@][/]:[:ro|:rw] ``` See 'docs/source/en/package_reference/hf_uris.md' for the full grammar and examples. N) dataclassfield)quoteunquoteurlsplit) constants) HfUriErrorHFValidationError)validate_repo_idci|]\}}|| Sr).0kvs b/var/www/app.qaans.com/backend/venv/lib/python3.11/site-packages/huggingface_hub/utils/_hf_uris.py r8s"["["[DAq1a"["["[_TYPE_TO_PREFIXz ^refs/(?:convert/[\w.-]+|pr/\d+)_VALID_URI_TYPES>rawblobtreeblameresolve_URL_REPO_LOCATION_ACTIONSrr_URL_BUCKET_LOCATION_ACTIONST)frozenceZdZUdZejed<eed<dZedzed<dZ eed<e dddd Z edzed <dd Z e d efd Ze d efdZd efdZddedzd efdZdS)HfUriaParsed representation of a Hugging Face Hub URI ('hf://...'). Attributes: type (`str`): One of 'model', 'dataset', 'space', 'kernel' or 'bucket'. id (`str`): The repository id ('namespace/name', e.g. 'my-org/my-model') for repo URIs, or the bucket id ('namespace/name') for bucket URIs. revision (`str`, *optional*): The revision specified after '@' in the URI, URL-decoded. 'None' if no revision was specified, or for bucket URIs (which never carry a revision). Special refs like 'refs/pr/10' and 'refs/convert/parquet' are preserved as-is. path_in_repo (`str`): The path inside the repo or bucket. Empty string if the URI points at the root. typeidNrevision path_in_repoFreprhashcomparedefault_rawreturnc |jpd}|jtvr/t|d|jdt td|jr|jddkrt|d|jd |jd krG t|jn1#t$r$}t|t||d}~wwxYw|j |j st|d |jd kr|j t|d |j r=|j ds d |j vrt|d|j ddSdS)Nr$zInvalid type 'z'. Must be one of .urimsg/r z"Id must be 'namespace/name', got ''.bucketz%Revision must not be an empty string.z&Bucket URIs do not support a revision.//z+Path must not contain empty segments (got 'z').) r+r!rr sortedr"countr r strr#r% startswith)selfr0es r __post_init__zHfUri.__post_init__dsio2 9, , ,*s49*s*sX^_oXpXp*s*s*sttt tw \$'--,,11*Ztw*Z*Z*Z[[[ [ 9  = ))))$ = = = Sc!ff5551< = = $T] $*QRRR R 9 T]%>*RSSS S   t ++C00 tDDr?s ris_repoz HfUri.is_reporArcrtjt|jd|jg}|jS|j}d|vr0t ||dd}| d||j r| d|j d |S)zRender the URI as a canonical 'hf://' string. The type prefix is always written explicitly (e.g. 'hf://models/my-org/my-model'). r2N%2F@r$) r HF_PROTOCOLrr!r"r#_SPECIAL_REFS_REVISION_REGEX fullmatchreplaceappendr%join)r:partsr#s rto_uriz HfUri.to_uris &1?493MsTXT[\ = $}Hh#?#I#I(#S#S#[#++C77 LLX ( ( (   2 LL0T.00 1 1 1wwu~~rendpointc|p tjd}t|jd}|jdkr|d|j}|r|d|z }|S|jdkr |d|jn|dt|jd|j}|j}|4d|vr0t || dd}|r|d |p tj d|z }n ||d|z }|S) aRender the URI as a Hugging Face **web URL** (the kind you open in a browser). This is the inverse of parsing a URL with [`parse_hf_uri`]. The returned URL points at: - the repository / bucket landing page when no path or revision is set; - the folder viewer ('/tree/') when only a revision is set; - the file viewer ('/blob//') for repository files (revision defaults to 'main'); - the tree route ('/tree/') for bucket files (buckets are not versioned). Args: endpoint (`str`, *optional*): Base endpoint to use. Defaults to 'constants.ENDPOINT' (i.e. 'https://huggingface.co'). Returns: `str`: the web URL. Example: ```py >>> from huggingface_hub import parse_hf_uri >>> parse_hf_uri("hf://datasets/my-org/my-dataset@v1/train.csv").to_url() 'https://huggingface.co/datasets/my-org/my-dataset/blob/v1/train.csv' ``` r2)safer4z /buckets/z/tree/modelNrEz/blob/) rENDPOINTrstriprr%r!r"rr#rHrIrJDEFAULT_REVISION)r:rObasepathurlr#s rto_urlz HfUri.to_urlsL0.I.66s;;T&S111 9 --DG--C '&J&*Y'%9%9!!!!!$?o?oY]YbIc?o?ofjfm?o?o=  C8OO8T8^8^_g8h8h8p''U33H  ' KHB (BKKTKK KCC  ! &H&& &C rr,NN)__name__ __module__ __qualname____doc__r HfUriType__annotations__r8r#r%rr+r<propertyboolr@rCrNrYrrrr r Ns1     GGGHcDjL#u%eUDQQQD#*QQQtttt6%4%%%X%%%%%X%$..sTz.S......rr c~eZdZUdZeed<eed<dZedzed<e ddddZ edzed<d d Z d efd Z dS) HfMountaFA HF URI paired with a local mount path and optional read-only flag. Used by Spaces and Jobs to describe volume mounts. The full syntax is: ``` hf://[/][@][/]:[:ro|:rw] ``` Attributes: source ([`HfUri`]): The parsed HF URI identifying the Hub resource to mount. mount_path (`str`): The local mount path (always starts with '/'). read_only (`bool`, *optional*): True if the mount ends with ':ro', False if it ends with ':rw', 'None' if no flag was provided. source mount_pathN read_onlyFr&r+r,c|jpd}|jdr |jdkrt|d|jddS)Nr$r2zEMount path must be a non-empty absolute path starting with '/', got 'r3r/)r+rgr9r )r:rs rr<zHfMount.__post_init__sfio2))#.. $/S2H2Ho\`\kooo 3I2Hrc|jd|jg}|j||jrdndd|S)zqRender the mount as a canonical 'hf://' string. Example: 'hf://models/my-org/my-model:/data:ro' :N:ro:rwr$)rfrNrgrhrKrL)r:rMs rrNzHfMount.to_urisU ##%%sDO< > % LL$.;e < < <wwu~~rrZ) r\r]r^r_r rar8rhrcrr+r<rNrrrreres" MMMOOO!Itd{!!!u%eUDQQQD#*QQQrrer0r,cH t|dS#t$rYdSwxYw)zVCheck if a string is a valid HF URI ('hf://...') or a recognized Hugging Face web URL.TF) parse_hf_urir )r0s r is_hf_urirps<St uus  !!rOc|}|tjrB|ttjd}|st |dtjdnNt ||rt ||}n+t |dtjdtjdt||\}}|d krt|||St|||S) aParse a Hugging Face Hub URI ('hf://...') or a Hugging Face web URL. A HF URI is a URI-like string identifying a location on the Hugging Face Hub. The full grammar is: ``` hf://[/][@][/] ``` For convenience, Hugging Face **web URLs** (the ones you copy-paste from the website) are also accepted and normalized to the canonical 'hf://' form, e.g. 'https://huggingface.co/datasets/my-org/my-dataset/blob/main/train.csv'. Only unambiguous URLs (repository / bucket pages and file/folder viewer routes) are accepted; any other route is rejected. See 'docs/source/en/package_reference/hf_uris.md' for the full specification. Args: uri (`str`): The URI to parse. Must start with 'hf://', or be a Hugging Face URL (e.g. 'https://huggingface.co/...'). endpoint (`str`, *optional*): A custom Hub endpoint (e.g. a self-hosted or proxied Hub like 'https://hub.my-company.com' or 'http://localhost:8080/hf'). When provided, web URLs on that endpoint are recognized in addition to the default Hugging Face hosts. Has no effect on 'hf://' URIs. Returns: [`HfUri`]: the parsed URI. Raises: [`HfUriError`]: If the URI is malformed (missing prefix, invalid type, missing id, unsupported URL route, etc.). Examples: ```py >>> from huggingface_hub.utils import parse_hf_uri >>> parse_hf_uri("hf://my-org/my-model") HfUri(type='model', id='my-org/my-model', revision=None, path_in_repo='') >>> parse_hf_uri("hf://datasets/my-org/my-dataset@refs/pr/3/train.json") HfUri(type='dataset', id='my-org/my-dataset', revision='refs/pr/3', path_in_repo='train.json') >>> parse_hf_uri("https://huggingface.co/datasets/my-org/my-dataset/blob/main/train.csv") HfUri(type='dataset', id='my-org/my-dataset', revision='main', path_in_repo='train.csv') ``` NEmpty body after 'r3)rOMust start with 'zQ' or be a Hugging Face URL (e.g. 'https://huggingface.co/...'). Expected format: z#[/][@][/]rr4) r9rrGlenr _looks_like_hf_url_url_to_uri_body _split_type_parse_bucket_body_parse_repo_body)r0rOrbodytype_locations rroros$V C ~~i+,,  3y,--//0 RS"Py7L"P"P"PQQ Q R C( 3 3 3 h777  [ 5 [ [ ) 5 [ [ [   "$C000OE8 !(Es;;;; He 5 5 55rc|dStd|vr|nd|z}|jr|jnd}||jdfS)aReturn the lowercased host and stripped path prefix of a custom Hub 'endpoint'. E.g. 'https://hub.my-company.com' -> ('hub.my-company.com', '') and a self-hosted 'http://localhost:8080/hf' -> ('localhost', 'hf'). Returns '(None, "")' when 'endpoint' is None. N)Nr$://r5r2)rhostnamelowerrWstrip)rOparsedhosts r_endpoint_host_and_pathr@sh x %8"3"3hh I IF&,o ?6? " " "4D ""3'' ''rcbt|\}}|rtj|hzn tjS)z^The set of hosts whose web URLs can be parsed: the default Hugging Face hosts plus 'endpoint'.)rr HF_URL_HOSTS)rOr_s r_recognized_hostsrNs0%h//GD!.2 N9 !TF * * 8NNrc|drdStfdt|DS)zNReturn True if 'uri' looks like a (possibly scheme-less) Hugging Face web URL.)zhttp://zhttps://Tc3TK|]"}|kp|dzV#dS)r2N)r9)rrlowereds r z%_looks_like_hf_url..Zs>iiTw$@'"4"4TCZ"@"@iiiiiir)rr9anyr)r0rOrs @rrvrvTsYiikkG122t iiiiM^_gMhMhiii i iirsegmentcHt|ddS)awPercent-decode a single URL path segment (e.g. 'file%20name.txt' -> 'file name.txt'). A decoded '/' is re-encoded as '%2F' so the segment stays atomic when the normalized body is re-split by the shared parser. This decodes ordinary path characters (spaces, '#', ...) that browsers encode, while keeping '%2F'-encoded revisions (e.g. 'feature%2Ffoo') intact. r2rE)rrJ)rs r_decode_url_path_segmentr]s" 7   # #C / //rrXcD|}td|vr|nd|z}|jpd}|t|vrt |d|p|d|j}t |\}}|r@||kr:d|z}||ks||dzr|t|d}d | dD} | st |d |d d} | d tj vr| d } | d d} t| dkrt |d|d| d d| d } | dd} | dkrZ| sd| S| ^} }| tvrt |d|d| dd d|D}|rd| d|nd| S| r| dnd}| s|| S| ^} }| tvrt |d|d| d|s|| Sd d|D}|| d|S)aNormalize a Hugging Face web URL into the body of a 'hf://' URI (everything after 'hf://'). The returned string is fed back into the regular URI parsing logic, so all validation (repo id, revision, empty path segments, ...) is shared with the canonical 'hf://' path. Only unambiguous URLs are accepted: any unrecognized route raises [`HfUriError`]. When 'endpoint' is provided, URLs on that custom Hub host are recognized too (and its path prefix is stripped). rr5r$zUnrecognized host 'zC'. Expected a Hugging Face URL (e.g. 'https://huggingface.co/...').r/r2Ncg|]}||Srrrrs r z$_url_to_uri_body..sBBBG'BBBBrz0Missing repository or bucket identifier in URL 'r3rr zCannot parse URL 'z{': expected a '/' repository or bucket. User/organization pages and single-segment URLs are not supported.bucketszbuckets/zCannot parse bucket URL 'z': unsupported '/z /' route.c34K|]}t|VdSr[rrs rrz#_url_to_uri_body..s+NNg099NNNNNNrzl/' route. Only repository pages and file/folder viewer routes (blob, resolve, raw, tree, ...) can be parsed.c34K|]}t|VdSr[rrs rrz#_url_to_uri_body..s+MMW/88MMMMMMrrF)rrrrr rWrr9rusplitrHF_URI_TYPE_PREFIXESrrLr)rXrOrrrrW endpoint_host endpoint_pathprefixsegments type_prefixrepo_idrestactiontaildecodeds rrwrwgsT C Uc\\cctcz : :F O !r ( ( * *D $X....vdkcvvv    ;D$;8#D#D M='..}$ 6>>T__Vc\::>F &DBBtzz#BBBH bS&`Y\&`&`&`aaaa#K{i444qk ABB< 8}}qUSUUU    !,,x{,,G ABB[:ro|:rw]'). A mount specification is a HF URI followed by a local mount path and an optional read-only/read-write flag. The full grammar is: ``` hf://[/][@][/]:[:ro|:rw] ``` See 'docs/source/en/package_reference/hf_uris.md' for the full specification. Args: mount_str (`str`): The mount string to parse. Must start with 'hf://' and contain a ':' segment. Returns: [`HfMount`]: the parsed mount. Raises: [`HfUriError`]: If the mount string is malformed (missing mount path, invalid URI, etc.). Examples: ```py >>> from huggingface_hub.utils import parse_hf_mount >>> parse_hf_mount("hf://my-org/my-model:/data:ro") HfMount(source=HfUri(type='model', id='my-org/my-model', revision=None, path_in_repo=''), mount_path='/data', read_only=True) >>> parse_hf_mount("hf://buckets/my-org/my-bucket/sub/dir:/mnt:rw") HfMount(source=HfUri(type='bucket', id='my-org/my-bucket', revision=None, path_in_repo='sub/dir'), mount_path='/mnt', read_only=False) ``` rsr3r/NrrrtzKMissing mount path. Expected ':' (e.g. 'hf://org/model:/data').)rfrgrhr+) r9rrGr ru _split_mountror1re) rrr{r}rgrhuri_strrfr;s rparse_hf_mountrs<@    5 6 6 =I$9===    C S.//11 2D VS&T9;P&T&T&TUUUU&24S&A&A&A#Hj)S&stttt#h.G4g&& 444Sae,,,!34 &Z9SV W W WWs;C C1C,,C1r{rc~|drd|d}}n/|drd|d}}nd}|d}|dkr|t|d |ddfS|d|}||d zd}|st|d |||fS) zSplit the ':[:ro|:rw]' suffix from 'body'. Returns '(location, mount_path, read_only)' where 'mount_path' is 'None' if no mount segment is present. rlTrmFNz:/zb':ro'/':rw' suffix is only valid when a mount path is provided (e.g. 'hf://...:/:ro').r/r z#Missing location before mount path.)endswith removesuffixrfindr )r{rrhidxr}rgs rrrs  }}U 1 1% 8 84 u  !2!25!9!94   **T  C byy  x T4DSDzHcAgiiJ MS&KLLLL Z **rr}c |d}|dkrs|tjvr$t|d|dtj|dt |x}!t|dtj|d d |fS|d|}||d zd}|tjvrtj||fSt |x}t|d |d |dd |fS)zDetect the (optional) type prefix and return '(type, remaining_location)'. A missing type prefix defaults to 'model'. Singular forms ('model/', 'dataset/', etc.) are explicitly rejected with a helpful error. r2rzMissing identifier after 'z '. Expected 'z/'.r/Nz*Type prefix must be plural. Did you mean 'z/...'?rRr z!Type prefix must be plural, got 'z/'. Did you mean 'z/'?)findrrr rGrget)r}r slash_idxsingular_pluralfirstrs rrxrxsV  c""IB y5 5 5ppp H]p_gppp  /228<< [@][/]'.r2zMissing repository id.r/rFrNr rz-Repository id must be 'namespace/name', got 'z'. rr$z!Missing repository id before '@'.r3zEmpty revision after '@'.r) rr rr7rrurHmatchgroup removeprefixrr ) r}r|rat_idxr#rMrr% rev_and_pathrrs rrzrzHsH~~c""H @S&>???? ]]3  F||x(..s33a77sA&& u::>>*gZb*g*g*ghhh h1X**a**#&u::>>uQxxr 7F7# -  O*MNNN N ==   " "*eZa*e*e*efff f-22<@@  {{}}H'H 8EEcJJLL$))#..IB'! '  3+IMOO< 8$$ G*EFFF F   !     rr[)0r_ functoolsre dataclassesrr urllib.parserrrhuggingface_hubrhuggingface_hub.errorsr r _validatorsr ritemsrdictr8racompilerH frozensetvaluesrrrr rercrp lru_cacherotuplerrrvrrwrrr`rxryrzrrrrsq: ((((((((1111111111%%%%%%@@@@@@@@)))))) #\"[I4R4X4X4Z4Z"["["[c3h[[[ *rz*MNN$-9Y-K-R-R-T-T#U#U)C.UUU .7Y7b7b7b-c-cIcNccc09y)V9L/M/MinMMM $zzzzzzzzz $''''''''T34 <6<6c<6S4Z<65<6<6<6<6~ (cDj (U3:s?5K ( ( ( (Od Oy~OOOO jjCj3:jjjjj0c0c0000O*O*#O*tO*sO*O*O*O*d7Xc7Xg7X7X7X7Xt+s+C+E#sTz4$;2N,O++++<#suY5H#5M/N>     088  8 8  888888r