^jk UdZddlZddlZddlmZmZddlmZmZm Z ddl m Z ddl m Z mZddlmZe j"j%Dcic]\}}|| c}}Zeeefed <ej.d Zee j"j5Zeeed <ehd Zeeed <eddhZeeed<edGddZedGddZdede fdZ!ejDd-dededzdefdZ#dedzde$edzeffdZ%dedzdeefdZ&d-dededzde fdZ'dedefd Z(d-d!ededzdefd"Z)d#edefd$Z*d%ed&ede$eedze dzffd'Z+d(ed&ede$e jXeffd)Z-d(ed*e jXd&edefd+Z.d(ed*e jXd&edefd,Z/ycc}}w).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_id_TYPE_TO_PREFIXz ^refs/(?:convert/[\w.-]+|pr/\d+)_VALID_URI_TYPES>rawblobtreeblameresolve_URL_REPO_LOCATION_ACTIONSrr_URL_BUCKET_LOCATION_ACTIONST)frozenceZdZUdZej ed<eed<dZedzed<dZ eed<e dddd Z edzed <dd Z e d efd Ze d efdZd efdZddedzd efdZy)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 |jxsd}|jtvr+t|d|jdt td|j r|j j ddk7rt|d|j d |jd k7r t|j |j|js t|d |jd k(r|j t|d |jrE|jjdsd |jvrt|d|jdyy#t$r}t|t||d}~wwxYw)NrzInvalid type 'z'. Must be one of .urimsg/r z"Id must be 'namespace/name', got ''.bucketz%Revision must not be an empty string.z&Bucket URIs do not support a revision.//z+Path must not contain empty segments (got 'z').) r#rrr sortedrcountr r strrr startswith)selfr(es i/var/www/ramen.bs-engineer-server.com/venv/lib/python3.12/site-packages/huggingface_hub/utils/_hf_uris.py __post_init__zHfUri.__post_init__ds]iio2 99, ,N499+EWX^_oXpWqqr*st tww$''--,1,NtwwiWY*Z[ [ 99  = ) == $T]]*QR R 99 T]]%>*RS S     ++C0DD') when only a revision is set; - the file viewer ('/blob//') for repository files (revision defaults to 'main'); - the tree route ('/tree/') for bucket files (buckets are not versioned). Args: endpoint (`str`, *optional*): Base endpoint to use. Defaults to 'constants.ENDPOINT' (i.e. 'https://huggingface.co'). Returns: `str`: the web URL. Example: ```py >>> from huggingface_hub import parse_hf_uri >>> parse_hf_uri("hf://datasets/my-org/my-dataset@v1/train.csv").to_url() 'https://huggingface.co/datasets/my-org/my-dataset/blob/v1/train.csv' ``` r*)safer,z /buckets/z/tree/modelrz/blob/) rENDPOINTrstriprrrrr rrBrCDEFAULT_REVISION)r2rIbasepathurlrs r4to_urlz HfUri.to_urls.0.I..66s;T&&S1 99 F)DGG9-Cv&J&*YY'%9ay!$qY]YbYbIcHddefjfmfmen?o==  $@$J$J8$T$\XB/H  VHB (B(BC1TFK KC  ! VH:& &C r;r$NN)__name__ __module__ __qualname____doc__r HfUriType__annotations__r0rrrr#r5propertyboolr9r=rHrSr;r4rrNs     GHcDjL#%eUDQD#*Qt6%4%%%%%$0sTz0S0r;rcveZdZUdZeed<eed<dZedzed<e ddddZ edzed<d d Z d efd Z y) HfMountaFA HF URI paired with a local mount path and optional read-only flag. Used by Spaces and Jobs to describe volume mounts. The full syntax is: ``` hf://[/][@][/]:[:ro|:rw] ``` Attributes: source ([`HfUri`]): The parsed HF URI identifying the Hub resource to mount. mount_path (`str`): The local mount path (always starts with '/'). read_only (`bool`, *optional*): True if the mount ends with ':ro', False if it ends with ':rw', 'None' if no flag was provided. source mount_pathN read_onlyFrr#r$c|jxsd}|jjdr|jdk(rt|d|jdy)Nrr*zEMount path must be a non-empty absolute path starting with '/', got 'r+r')r#rbr1r )r2rs r4r5zHfMount.__post_init__sWiio2))#.$//S2H[\`\k\k[llno 3Ir;c|jjd|jg}|j|j |jrdnddj |S)zqRender the mount as a canonical 'hf://' string. Example: 'hf://models/my-org/my-model:/data:ro' ::ro:rwr)rarHrbrcrErF)r2rGs r4rHzHfMount.to_urisM ##%sDOO< >> % LL$..e <wwu~r;rT) rVrWrXrYrr[r0rcr]rr#r5rHr^r;r4r`r`sK" MO!Itd{!%eUDQD#*Qr;r`r(r$c: t|y#t$rYywxYw)zVCheck if a string is a valid HF URI ('hf://...') or a recognized Hugging Face web URL.TF) parse_hf_urir )r(s r4 is_hf_urirks$S s  rIc|}|jtjr<|ttjd}|sht |dtjdt ||rt ||}n/t |dtjdtjdt||\}}|d k(rt|||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 'r+)rIMust start with 'zQ' or be a Hugging Face URL (e.g. 'https://huggingface.co/...'). Expected format: z#[/][@][/]rr,) r1rrAlenr _looks_like_hf_url_url_to_uri_body _split_type_parse_bucket_body_parse_repo_body)r(rIrbodytype_locations r4rjrjsV C ~~i++,3y,,-/0S$6y7L7L6MR"PQ Q C( 3h7  5 567 ) 5 566Y [  "$C0OE8 !(Es;; He 55r;c|ytd|vr|nd|z}|jr|jjnd}||jj dfS)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://r-r*)rhostnamelowerrQstrip)rIparsedhosts r4_endpoint_host_and_pathrBsX  %8"3h IF&,oo6?? "4D ""3' ''r;cjt|\}}|rtj|hzStjS)z^The set of hosts whose web URLs can be parsed: the default Hugging Face hosts plus 'endpoint'.)rr HF_URL_HOSTS)rIr_s r4_recognized_hostsrPs0%h/GD!.29 ! !TF *N 8N8NNr;c|jjdrytfdt|DS)zNReturn True if 'uri' looks like a (possibly scheme-less) Hugging Face web URL.)zhttp://zhttps://Tc3TK|]}|k(xsj|dz!yw)r*N)r1).0rlowereds r4 z%_looks_like_hf_url..\s,iTw$@'"4"4TCZ"@@is%()r|r1anyr)r(rIrs @r4rqrqVs8iikG12 iM^_gMhi iir;segmentc8t|jddS)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. r*r?)rrD)rs r4_decode_url_path_segmentr_s 7  # #C //r;rRc|}td|vr|nd|z}|jxsdj}|t|vrt |d|xs|d|j }t |\}}|r1||k(r,d|z}||k(s|j|dzr|t|d}|jdD cgc]} | s|  } } | st |d |d d} | d tjvr | d } | d d} t| d krt |d|d| d d| d } | d d} | dk(rN| sd| S| ^}}|tvrt |d|d|ddjd|D}|rd| d|Sd| S| r| dnd}| s|| S| ^}}|tvrt |d|d|d|s|| Sdjd|D}|| d|Scc} w)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). rzr-rzUnrecognized host 'zC'. Expected a Hugging Face URL (e.g. 'https://huggingface.co/...').r'r*Nz0Missing repository or bucket identifier in URL 'r+rr 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.c32K|]}t|ywrUrrrs r4rz#_url_to_uri_body..sNg09Nzl/' route. Only repository pages and file/folder viewer routes (blob, resolve, raw, tree, ...) can be parsed.c32K|]}t|ywrUrrs r4rz#_url_to_uri_body..sMW/8Mrr@)rr{r|rr rQrr1rpsplitrHF_URI_TYPE_PREFIXESrrFr)rRrIrr~rrQ endpoint_host endpoint_pathprefixrsegments type_prefixrepo_idrestactiontaildecodeds r4rrrris C Uc\ctcz :F OO !r ( ( *D $X..%dkc]2uv  ;;D$;8#D M=.}$ 6>T__Vc\:F &D'+zz#BG'BHB S(XY\X]]_&`aa#K{i444qk AB< 8}q$SE*UU  ! Qx{m,G AB[: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) ``` rnr+r'NrmrozKMissing mount path. Expected ':' (e.g. 'hf://org/model:/data').)rarbrcr#) r1rrAr rp _split_mountrjr)r`) rrrvrxrbrcuri_strrar3s r4parse_hf_mountrs@    5 5 6#I$9$9#:"=  C S../1 2D S(:9;P;P:QQS&TUU&24S&A#Hj)S&stt##h.G4g& &Z9SV WW 4Saee,!34s2 C C2C--C2rvrc0|jdrd|jd}}n'|jdrd|jd}}nd}|jd}|dk(r| t|d |ddfS|d|}||d zd}|s t|d |||fS) zSplit the ':[:ro|:rw]' suffix from 'body'. Returns '(location, mount_path, read_only)' where 'mount_path' is 'None' if no mount segment is present. rgTrhFNz:/zb':ro'/':rw' suffix is only valid when a mount path is provided (e.g. 'hf://...:/:ro').r'r z#Missing location before mount path.)endswith removesuffixrfindr )rvrrcidxrxrbs r4rrs  }}U 1 1% 84 u !2!25!94   **T C by  x T4DSzHcAgiJ S&KLL Z **r;rxc |jd}|dk(rr|tjvr$t|d|dtj|dt j |x}!t|dtj|d d |fS|d|}||d zd}|tjvrtj||fSt j |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. r*rzMissing identifier after 'z '. Expected 'z/'.r'Nz*Type prefix must be plural. Did you mean 'z/...'?rLr z!Type prefix must be plural, got 'z/'. Did you mean 'z/'?)findrrr rAr get)rxr slash_idxsingular_pluralfirstrs r4rsrss/  c"IB y55 50 - H]H]G^_g^hhop  /228< [@][/]'.r*zMissing repository id.r'r@rNr rz-Repository id must be 'namespace/name', got 'z'. rrz!Missing repository id before '@'.r+zEmpty revision after '@'.r) r}r rr/rrprBmatchgroup removeprefixrr) rxrwrat_idxrrGrr rev_and_pathrrs r4ruruJs~~c"H S&>?? ]]3 F|x(..s3a7sA& u:>,YZbYccf*gh h1XJaaz*#&u:>uQxr 7F# - *MN N ==  ",YZaYbbd*ef f-22<@  {{}H'H 8EEcJL$))#.IB'! ' 3+IMO< 8$*EF F   !   r;rU)0rY functoolsre dataclassesrr urllib.parserrrhuggingface_hubrhuggingface_hub.errorsr r _validatorsr ritemsr dictr0r[compilerB frozensetvaluesrrrrr`r]rk lru_cacherjtuplerrrqrrrrrrZrsrtru)kvs00r4rs: (11%@) 5>4R4R4X4X4Z"[DAq1a4"[c3h[ *rzz*MN$-Y-K-K-R-R-T#U)C.U .77b-cIcNc09)V9L/MinM $|||~ $'''T34 <6c<6S4Z<65<6<6~ (cDj (U3:s?5K (Od Oy~O jCj3:jj0c0c0O*#O*tO*sO*d7Xc7Xg7Xt+s+C+E#sTz4$;2N,O+<#suY5H5H#5M/N>      088   8 8  8e#\s G