*pj/%SrSSKJr SSKrSSKJr SSKJr 0S\"SS15_S \"S S 15_S \"S 15_S \"S15_S\"S15_S\"SS15_S\"SS15_S\"S15_S\"S15_S\"S15_S\"S15_S\"S15_S\"1Sk5_S \"S!S"15_S#\"S$15_S%\"S15_S&\"S'15_\"S(15\"S)15\"S*15\"S+15\"S,15\"S-15\"S.S/15\"S015\"S015\"S1S215\"S315\"S315S4. Er S5\ S6'S7r S8\ S9'S:r S8\ S;'S<4SNS=jjr \ "\ 5rS5\ S>'SOS?jr0rS5\ S@'\ R#5HCurr\"\5=(d \r\"SA\55r\R-\\"55\-\\'ME \ "\ SBSC9rS5\ SD'SPSEjrSPSFjrSQSGjr\"SHSI15rSJ\ SK'SRSLjrSSSMjrg)TaAccuracy-evaluation tables and predicates. Tools for asking *"is this detection acceptable, given what was expected?"* Used by the test suite and by benchmarking/diagnostic scripts (e.g., ``scripts/compare_detectors.py``, ``scripts/diagnose_accuracy.py``). This module owns three kinds of tables: 1. **Directional supersets** (:data:`SUPERSETS`) -- detecting a superset encoding when the expected encoding is a subset is correct (e.g., detecting UTF-8 when expected is ASCII), but not the reverse. 2. **Bidirectional encoding groups** (:data:`BIDIRECTIONAL_GROUPS`) -- groups of encodings where any member is interchangeable with any other. 3. **Bidirectional language groups** (:data:`LANGUAGE_EQUIVALENCES`) -- ISO 639-1 codes for languages that are nearly indistinguishable by statistical detection (e.g., Slovak/Czech, Malay/Indonesian). The corresponding predicates are :func:`is_correct`, :func:`is_equivalent_detection`, and :func:`is_language_equivalent`. ) annotationsN)Callablelookup_encodingASCIIzutf-8cp1252zTIS-620z iso8859-11cp874z ISO-8859-11GB2312gb18030GBKBig5 big5hkscscp950 Shift_JIScp932shift_jis_2004zShift-JISX0213zEUC-JP euc_jis_2004z EUC-JISX0213zEUC-KRcp949CP037cp1140z ISO-2022-JP> iso2022_jp_2iso2022_jp_extiso2022_jp_2004z ISO2022-JP-1rrz ISO2022-JP-3rz ISO-8859-1z ISO-8859-2cp1250cp1251cp1256cp1253cp1255cp1254cp1257z utf-16-lez utf-16-bezutf-16z utf-32-lez utf-32-bezutf-32) z ISO-8859-5z ISO-8859-6z ISO-8859-7z ISO-8859-8z ISO-8859-9z ISO-8859-13zUTF-16z UTF-16-LEz UTF-16-BEzUTF-32z UTF-32-LEz UTF-32-BEdict[str, frozenset[str]] SUPERSETS))rrrtuple[tuple[str, ...], ...]BIDIRECTIONAL_GROUPS))skcs)ukrubgbe)msid)nodasvLANGUAGE_EQUIVALENCEScU$N)xs N/var/www/html/pdf-tiff/venv/lib/python3.13/site-packages/chardet/evaluation.pyr6wscn^0nUH+n[U4SjU55nUH nXBT"U5'M M- U$)zJBuild a lookup: key -> frozenset of all equivalent keys in the same group.c34># UH nT"U5v M g7fr2r3).0n normalizes r5 %_build_group_index..|s7A9Q<<s) frozenset)groupsr<resultgroupnormednames ` r5_build_group_indexrEusA )+F777D&,9T? # Mr7_LANGUAGE_EQUIVcVX:Xag[RU5nUSL=(a X;$)avCheck whether *detected* is an acceptable language for *expected*. Returns ``True`` when *expected* and *detected* are the same ISO 639-1 code, or belong to the same equivalence group in :data:`LANGUAGE_EQUIVALENCES`. :param expected: Expected ISO 639-1 language code. :param detected: Detected ISO 639-1 language code. :returns: ``True`` if the languages are equivalent. TN)rFget)expecteddetectedrBs r5is_language_equivalentrKs0    )E   2!22r7_NORMALIZED_SUPERSETSc#J# UHn[U5=(d Uv M g7fr2r)r:ss r5r=r=sDA*/a/s!#c*[U5=(d U$r2r)r;s r5r6r6soa.@.EA.Er7)r<_NORMALIZED_BIDIRcUcUSL$Ucg[U5=(d UR5n[U5=(d UR5nX#:H$)aCheck whether *detected* names the very same encoding as *expected*. Only alias and case differences are normalized away, so ``windows-1252`` matches ``cp1252``. Unlike :func:`is_correct`, this credits neither supersets nor byte-order variants: detecting UTF-8 for an ASCII file is ``False`` here. This is the stricter convention some other detectors score themselves against. Reporting it alongside :func:`is_correct` makes the two conventions directly comparable instead of leaving the difference implicit. :param expected: The expected encoding name, or ``None`` for binary files. :param detected: The detected encoding name, or ``None``. :returns: ``True`` if both names resolve to the same encoding. NF)rlowerrIrJnorm_expnorm_dets r5is_exact_matchrVsS"4x(.sCd+*?*?*B11ds. .)r]r<join)textnfkds r5_strip_combiningrds)   .D 77CdC CCr7)¤€)rfrezfrozenset[tuple[str, str]]_EQUIVALENT_SYMBOL_PAIRScTX:XagX4[;ag[U5[U5:H$)uReturn True if characters *a* and *b* are functionally equivalent. Equivalent means: - Same character, OR - Same base letter after stripping combining marks, OR - An explicitly listed symbol equivalence (e.g. ¤ ↔ €) T)rgrd)abs r5_chars_equivalentrks0 v v)) A "21"5 55r7c UcUSL$Ucg[U5=(d UR5n[U5=(d UR5nX4:XagURU5nURU5nXV:Xag[ U5[ U5:wag[ S[XVSS955$![[4a gf=f)uUCheck whether *detected* produces functionally identical text to *expected*. Returns ``True`` when: 1. *detected* is not ``None`` and both encoding names normalize to the same codec, OR 2. Decoding *data* with both encodings yields identical strings, OR 3. Every differing character pair is functionally equivalent: same base letter after stripping combining marks, or an explicitly listed symbol equivalence (e.g. ¤ ↔ €). Returns ``False`` if *detected* is ``None``, either encoding is unknown, or either encoding cannot decode *data*. :param data: The raw byte data that was detected. :param expected: The expected encoding name, or ``None`` for binary files. :param detected: The detected encoding name, or ``None``. :returns: ``True`` if decoding with *detected* yields functionally identical text to decoding with *expected*. NFTc3<# UHup[X5v M g7fr2)rk)r:rirjs r5r=*is_equivalent_detection..4sX3W41 &&3Ws)strict)rrRdecodeUnicodeDecodeError LookupErrorlenallzip)datarIrJrTrUtext_exptext_dets r5is_equivalent_detectionrys.4x(rsB.#$,+( Y* ++( y,01+(9gY'+( i $ +(  9i[ ! +(  I{G, - +(G%567+(i!1 23+( i()+(I~./+( i "+( Yz "+(.9RS/+(0I~/?@A1+(2I0123+(6)XJ'7+(8)XJ'9+(:XJ'XJ'XJ'XJ'XJ'hZ(k23H:&H:&k23H:&H:&U+( $+h5162&'2  ' #  .@@U-V*V3*4605$??,GZ 7 # .wDDDDG"7";";D)+"NQX"X$- 0B$E0,  4#LD8A84 6 .Y .Y%.Y1;.Y .Yr7