FjdZddlZddlmZmZmZmZddlmZddddddd d d Z d e d e fdZ d@de de de de d ee e ee ee ff dZde deee e fde de d ee f dZde d e fdZd e d ee fdZde de de d e fdZde de deee e fd e fdZ dAde deee e fde dee d e f dZde de d eee e ffdZde de d eee e ffd Zde de d eee e ffd!Zde de d eee e ffd"Zde de d eee e ffd#Zde de d eee e ffd$Zd%e d ee fd&Zd'ee d(eee e fd eee e ffd)Zde de d eee e ffd*Zde de d eee e ffd+Zde de d eee e ffd,Zd-ee d.e d/e d0e d ee e ff d1Z de d-ee d2ee de d3e d eee e ff d4Z!d%e d5e d6eee e fd eee e ffd7Z"dBde de d:e d;e d e f d<Z#d=ee d>e de de d e f d?Z$dS)CaP Fuzzy Matching Module for File Operations Implements a multi-strategy matching chain to robustly find and replace text, accommodating variations in whitespace, indentation, and escaping common in LLM-generated code. The 8-strategy chain (inspired by OpenCode), tried in order: 1. Exact match - Direct string comparison 2. Line-trimmed - Strip leading/trailing whitespace per line 3. Whitespace normalized - Collapse multiple spaces/tabs to single space 4. Indentation flexible - Ignore indentation differences entirely 5. Escape normalized - Convert \n literals to actual newlines 6. Trimmed boundary - Trim first/last line whitespace only 7. Block anchor - Match first+last lines, use similarity for middle 8. Context-aware - 50% line similarity threshold Multi-occurrence matching is handled via the replace_all flag. Usage: from tools.fuzzy_match import fuzzy_find_and_replace new_content, match_count, strategy, error = fuzzy_find_and_replace( content="def foo():\n pass", old_string="def foo():", new_string="def bar():", replace_all=False ) N)TupleOptionalListCallable)SequenceMatcher"'z---z... )u“u”u‘u’u—u–u… textreturncptD]\}}|||}|S)zBNormalizes Unicode characters to their standard ASCII equivalents.) UNICODE_MAPitemsreplace)r charrepls 0/usr/local/lib/hermes-agent/tools/fuzzy_match.py_unicode_normalizer+s;!''))(( d||D$'' KFcontent old_string new_string replace_allc |s|dddfS||kr|dddfSdtfdtfdtfdtfd tfd t fd t fd tfd tfg }|D]\}}|||}|rt|dkr|s|dddt|dfcS|dkrt||||}|r|dd|fcSt|||} t||| |dkr|nd} | t||dfcS|dddfS)a) Find and replace text using a chain of increasingly fuzzy matching strategies. Args: content: The file content to search in old_string: The text to find new_string: The replacement text replace_all: If True, replace all occurrences; if False, require uniqueness Returns: Tuple of (new_content, match_count, strategy_name, error_message) - If successful: (modified_content, number_of_replacements, strategy_used, None) - If failed: (original_content, 0, None, error_description) rNzold_string cannot be emptyz'old_string and new_string are identicalexact line_trimmedwhitespace_normalizedindentation_flexibleescape_normalizedtrimmed_boundaryunicode_normalized block_anchor context_awarezFound zY matches for old_string. Provide more context to make it unique, or use replace_all=True.)rz1Could not find a match for old_string in the file) _strategy_exact_strategy_line_trimmed_strategy_whitespace_normalized_strategy_indentation_flexible_strategy_escape_normalized_strategy_trimmed_boundary_strategy_unicode_normalized_strategy_block_anchor_strategy_context_awarelen_detect_escape_drift_maybe_unescape_new_string_apply_replacements) rrrr strategies strategy_name strategy_fnmatches drift_err effective_new new_contents rfuzzy_find_and_replacer;2s >4!===Z4!JJJ /" /0 "AB !?@ 9: 78 ;< /0 12 .J'18B8B" {+gz22 5 B7||a 4XS\\XXX''0':zZZ 7"AtY666627GWM.-)6')A)A::tKG mTA A A Ak5 Bp AtP PPrr7cd|vrd|vrdSdfd|D}dD]!}||vr||vr||vr|d}d|d |d cS"dS) u'Detect tool-call escape-drift artifacts in new_string. Looks for ``\'`` or ``\"`` sequences that are present in both old_string and new_string (i.e. the model copy-pasted them as "context" it intended to preserve) but don't exist in the matched region of the file. That pattern indicates the transport layer inserted spurious shell-style escapes around apostrophes or quotes — writing new_string verbatim would literally insert ``\'`` into source code. Returns an error string if drift is detected, None otherwise. \'\"Nc32K|]\}}||VdSN.0startendrs r z'_detect_escape_drift..0KKZUCgeCi0KKKKKKr)r=r>r&zNEscape-drift detected: old_string and new_string contain the literal sequence a but the matched region of the file does not. This is almost always a tool-call serialization artifact where an apostrophe or quote got prefixed with a spurious backslash. Re-read the file with read_file and pass old_string/new_string without backslash-escaping z characters.)join)rr7rrmatched_regionssuspectplains` rr1r1s  J5 #:#:t ggKKKK7KKKKKO!   j W %:%:wo?]?]AJE<(/<< ', <<<    4rlinecd}|t|kr,||dvr"|dz }|t|kr ||dv"|d|S)z=Return the leading whitespace prefix of a line (spaces/tabs).r)r  r&Nr0)rMis r_leading_whitespacerRsY A c$ii--DG{22 Q c$ii--DG{22 8Orcf|dD]}|r|cSdS)zReturn the first line of ``text`` that has any non-whitespace content. Returns ``None`` if no such line exists (text is empty or all whitespace).  N)splitstrip)r rMs r_first_meaningful_linerWsB  4   ::<< KKK  4r file_regioncN|s|St|}t|}|||St|}t|}||kr|Sg}|dD]}|s||,t|} | |r0|t |d} ||| z|||dzd|S)a[Adjust ``new_string`` so its indentation matches ``file_region``. Used after a non-exact fuzzy match: the LLM may have sent old_string and new_string with a different indent than the file actually has (e.g. 2-space indent in tool args vs 4-space indent on disk). The fuzzy strategy successfully matched anyway, but writing ``new_string`` verbatim would corrupt the file's indentation. Approach: 1. For each non-blank line in ``new_string``, compute its indent *relative* to the shallowest non-blank line of ``old_string`` (the LLM's base indent). 2. Anchor that relative indent onto the file's actual base indent (the leading whitespace of the file_region's first non-blank line). 3. Re-emit each non-blank line as ``file_base + (line_indent - llm_base)``. Blank lines and lines less-indented than the LLM's base are anchored directly to the file's base indent. No-op cases (returns ``new_string`` unchanged): - file_region or old_string has no meaningful line - LLM base indent equals file base indent - new_string is empty NrT ) rWrRrUrVappend startswithr0lstriprI) rXrr old_first file_first old_indent file_indent out_linesrM line_indent remainders r_reindent_replacementresF4 &z22I' 44JJ.$Y//J%j11K[  I  &&??zz||    T " " " )$//  ! !* - - ?S__--.I   [94 5 5 5 5   [4;;u+=+== > > > > 99Y  rcd|vrd|vr|Sdfd|D}|}d|vrd|vr|dd}d|vrd|vr|dd}|S)u Conditionally unescape ``\t``/``\r`` in new_string. LLMs frequently send the two-character sequences ``\t`` (backslash + t) and ``\r`` (backslash + r) inside JSON tool-call arguments where they meant a real tab or carriage-return byte. Writing the string verbatim corrupts tab-indented files with literal backslash-letter pairs. The unescape is only applied per-sequence when the *matched region of the file* actually contains the corresponding control character — that is, we only convert ``\t`` -> tab when the file region we're replacing contains a real tab byte. Files that legitimately contain the literal two-character string ``"\t"`` (e.g. a Python source line that defines ``sep = "\t"``) get a backslash+t in the matched region instead of a tab, so we leave new_string alone. ``\n`` is intentionally excluded: newlines serialize correctly through JSON and rewriting backslash-n would corrupt escape sequences in string literals far more often than it would help. \t\rr?c32K|]\}}||VdSrArBrCs rrGz-_maybe_unescape_new_string..*rHrrO )rIr)rrr7rJouts ` rr2r2s0 J5 #:#:ggKKKK7KKKKKO C ||//kk%&& ||//kk%&& Jrct|dd}|}|D]=\}}||||}t|||} n|} |d|| z||dz}>|S)a Apply replacements at the given positions. Args: content: Original content matches: List of (start, end) positions to replace new_string: Replacement text old_string: When non-None, signals that the match came from a non-exact fuzzy strategy; ``new_string`` is re-indented to match the file's actual indentation before substitution. Returns: Content with replacements applied c|dSNrrBxs rz%_apply_replacements..Es 1Q4rT)keyreverseN)sortedre) rr7rrsorted_matchesresultrErFrXadjusteds rr3r33s$GFFFN F$:: s  !!%),K,[*jQQHH!H(*VCDD\9 Mrpatterncg}d} |||}|dkrn-|||t|zf|dz}J|S)zStrategy 1: Exact string match.rTr&)findr[r0)rrxr7rEposs rr'r'WsgG Ell7E** "99 S3w<</0111a  Nrcd|dD}d|}|d}d|D}t|||||S)z Strategy 2: Match with line-by-line whitespace trimming. Strips leading/trailing whitespace from each line before matching. c6g|]}|SrBrVrDrMs r z*_strategy_line_trimmed..ks BBBdTZZ\\BBBrrTc6g|]}|SrBrrs rrz*_strategy_line_trimmed..os GGG GGGr)rUrI_find_normalized_matches)rrx pattern_linespattern_normalized content_linescontent_normalized_liness rr(r(ds}CBgmmD.A.ABBBM=11MM$''MGGGGG $ 8#  rc~d}||}||}t||}|sgSt|||S)zC Strategy 3: Collapse multiple whitespace to single space. c.tjdd|S)Nz[ \t]+r )resubss r normalizez2_strategy_whitespace_normalized..normalize|svia(((r)r'_map_normalized_positions)rrxrrcontent_normalizedmatches_in_normalizeds rr)r)xsj)))#7++"7++,,>@RSS   %W.@BW X XXrc |d}d|D}d|dD}t||||d|S)z Strategy 4: Ignore indentation differences entirely. Strips all leading whitespace from lines before matching. rTc6g|]}|SrBr]rs rrz2_strategy_indentation_flexible..s FFFdkkmmFFFrc6g|]}|SrBrrs rrz2_strategy_indentation_flexible..s CCCtT[[]]CCCr)rUrrI)rrxrcontent_stripped_linesrs rr*r*ss MM$''MFF FFFCCw}}T/B/BCCCM # 6=))  rcNd}||}||krgSt||S)zt Strategy 5: Convert escape sequences to actual characters. Handles \n -> newline, \t -> tab, etc. c~|ddddddS)Nz\nrTrgrOrhrj)rrs runescapez-_strategy_escape_normalized..unescapes6yy%%--eT::BB5$OOOr)r')rrxrpattern_unescapeds rr+r+sG PPP!))G## 7$5 6 66rc |d}|sgS|d|d<t|dkr|d|d<d|}|d}g}t|}t t||z dzD]}||||z}|} | d| d<t| dkr| d| d<d| |kr ? ?11#Aa*<&<$<= "&&(( $Q--// A {  a  )"o3355KO 99[ ! !%5 5 5!:q!&8"8#g,,"" Iw NNIw/ 0 0 0 Nroriginalcg}d}|D]G}||t|}||t|ndz }H|||S)uBuild a list mapping each original character index to its normalized index. Because UNICODE_MAP replacements may expand characters (e.g. em-dash → '--', ellipsis → '...'), the normalised string can be longer than the original. This map lets us convert positions in the normalised string back to the corresponding positions in the original string. Returns a list of length ``len(original) + 1``; entry ``i`` is the normalised index that character ``i`` maps to. rNr&)r[rgetr0)rrvnorm_posrrs r_build_orig_to_norm_maprsrFH99 ht$$!1CIIIq8 MM( Mr orig_to_norm norm_matchesc,i}t|ddD]\}}||vr|||<g}t|dz }|D]T\}}||vr ||} | } | |kr#|| |kr| dz } | |kr || |k|| | fU|S)zNConvert (start, end) positions in the normalised string to original positions.Nrzr&) enumerater0r[) rrnorm_to_orig_startorig_posrresultsorig_len norm_startnorm_end orig_startorig_ends r_map_positions_norm_to_origrs *,' SbS(9::44( - - -+3 x (%'G<  1$H , / / H / / / ' 3 !!l8&>L J-lLII  *733L &|\ B BBrc t|}t|}|d}t|dkrgS|d}|d}|d}|d}t|} g} t t|| z dzD]Y} || |kr9|| | zdz |kr| | Zg} t| } | dkrdnd}| D]} | dkrd}nfd|| dz| | zdz }d|dd}td ||}||krRST>T:T'UVVN!YY}QrT':;;N(~~NNTTVVJ  " "!:"Aq+='=s7||"" Iw NNIw/ 0 0 0 Nrc `|d}|d}|sgSg}t|}tt||z dzD]}||||z}d}t||D]W\} } t d| | } | dkr|dz }X|t|dzkr ? ?11#Aa*<&<$<= !"!-== + +NFF!$  GGMMOOCd{{%*% !C $6$6$< < .s,IIdCIIMIIIIIIrNc3:K|]}t|dzVdSrrPrs rrGz,_calculate_line_positions..s,EED#d))a-EEEEEErr&)summin)rrrrrrs rrrsmIImKZK.HIIIIIIEEM)8),DEEEEEIG.'**G g rrrc `|d}t|}g}tt||z dzD]d}d||||z} | |kr.s+VV41aa:ooQooooVVr)r0r[rr)rrrrorig_idxnorm_idxrnorm_to_orig_endrroriginal_matchesrrrrs @rrrs   LHH S]] " "x#j//'A'A H H!5 5 5    ) ) ) MH MHH h 5 ( (Z-AS-H-H    ) ) ) MH#h--''HX,>e,K,KA  h 5 ( (    ) ) ) MHH    ) ) ) MH' S]] " "x#j//'A'A, S]] " "C OO,,,A  S]] " " ' 55..( - - -+3 x (%-"" 2LL H + + ++J7JJVVVV9\+B+BVVVVVJ a<+ + +'1 59HH!X %:;HX&&8H+=+F+F MHX&&8H+=+F+F S3x==-I-I JKKKK rr context_lines max_resultsc|r|sdS|}||rsdS|d}|sd|D}|sdS|d}g}tD]\\}} | } | std|| } | dkr|| |f]|sdS|d|d|} g} t}| D]\}}td||z tt|t|z|z}|f}||vrQ| |d fd t|z D}| || sdSd | S) zFind lines in content most similar to old_string for "did you mean?" feedback. Returns a formatted string showing the closest matching lines with context, or empty string if no useful match is found. r?rc^g|]*}||+SrBr)rDls rrz&find_closest_lines..s-@@@Aaggii@aggii@@@rNg333333?c|d SrnrBros rrqz$find_closest_lines..2s qter)rrrTc3FK|]}|zdzdd|zVdS)r&4dz| NrB)rDjrrEs rrGz%find_closest_lines..>sY  qy1} = = ==#; = =      rz --- ) splitlinesrVrrrr[sortsetmaxrr0addrIr)rrrr old_linesanchor candidatesscoredrQrMstrippedrtopparts seen_ranges_line_idxrFrrsnippetrrEs @@rfind_closest_linesr sG Wr%%''I&&((M Mrq\   ! !F @@@@@  2AF]++&&4::<<  fh77==?? 3;; MM5!* % % % r KKOOK$$$ + C E%%K   8Ax-/00#m$$hY&?-&OPPcl +   ))     3;''       W r >>%  rerror match_countcv|dkrdS|r|dsdSt||}|sdSd|zS)uReturn a '\n\nDid you mean...' snippet for plain no-match errors. Gated so the hint only fires for actual "old_string not found" failures. Ambiguous-match ("Found N matches"), escape-drift, and identical-strings errors all have ``match_count == 0`` but a "did you mean?" snippet would be misleading — those failed for unrelated reasons. Returns an empty string when there's nothing useful to append. rr?zCould not findz& Did you mean one of these sections? )r\r)rrrrhints rformat_no_match_hintrJs^ar (()9::r j' 2 2D r 6 ==r)FrA)rr)%__doc__rtypingrrrrdifflibrrstrrboolintr;r1rRrWrer2r3r'r(r)r*r+r,rrr-r.r/rrrrrrBrrr sT< 222222222222######SScs  SS16^Q^QC^QS^Qc^Q)-^Q:?S(SV-YabeYf@f:g^Q^Q^Q^QB%#%U38_0E%%(%69%>Fsm%%%%Pcc#> s> > > QT> > > > B!3!(+!(,U38_(=!BE!!!!JFJtE#s(O/D$'5=c]NQH S 3 4c3h3H    C#$uS#X:O(YSY3Y4cSVhCXYYYY* C # $uSRUXBW     77s7tE#s(O?T7777&''c'd5c?>S''''Tcd3i*s)uS#X' %S/:C#CCU3PS8_@UCCCC>5C5#5$uS#X:O5555p S 3 4c3h;P    NT#YC(+=@EJ3PS8_& c $s) 8 >C>%(>36>;>>>>>>>r