[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7BYZEBPDD7SN7TC3FVY4HG7":3,"skill-same:01M7BYZEBPDD7SN7TC3FVY4HG7":45},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":24,"commitSha":25,"skillDirectory":26,"forksCount":27,"repositoryHtmlUrl":28,"repositoryHomepageUrl":6,"ownerHtmlUrl":29,"ownerBlogUrl":6,"ownerDescription":6,"updatedAt":30,"install":31,"resources":44},"01M7BYZEBPDD7SN7TC3FVY4HG7","fuzzy-match",null,"A toolkit for fuzzy string matching and data reconciliation. Useful for matching entity names (companies, people) across different datasets where spelling variations, typos, or formatting differences exist.","linyouxue\u002Fskill-repair-benchmark","linyouxue","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F216151464?v=4","01M7BJEVCXKD74KH6FG9CBVSQB","林宥学","01M7BJEVE02NP217V93BKZFC0J",0,"MIT","https:\u002F\u002Fgithub.com\u002Flinyouxue\u002Fskill-repair-benchmark\u002Fblob\u002Fmain\u002Fevaluation\u002Fresults\u002FSkillAxe\u002Ftasks\u002Finvoice-fraud-detection\u002Frepaired_run\u002Finputs\u002Fskills\u002Ffuzzy-match\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: fuzzy-match\ndescription: A toolkit for fuzzy string matching and data reconciliation. Useful for matching entity names (companies, people) across different datasets where spelling variations, typos, or formatting differences exist.\nlicense: MIT\n---\n\n# Fuzzy Matching Guide\n\n## Overview\n\nThis skill provides methods to compare strings and find the best matches using Levenshtein distance and other similarity metrics. It is essential when joining datasets on string keys that are not identical.\n\n## Quick Start\n\n```python\nfrom difflib import SequenceMatcher\n\ndef similarity(a, b):\n    return SequenceMatcher(None, a, b).ratio()\n\nprint(similarity(\"Apple Inc.\", \"Apple Incorporated\"))\n# Output: 0.7...\n```\n\n## Python Libraries\n\n### difflib (Standard Library)\n\nThe `difflib` module provides classes and functions for comparing sequences.\n\n#### Basic Similarity\n\n```python\nfrom difflib import SequenceMatcher\n\ndef get_similarity(str1, str2):\n    \"\"\"Returns a ratio between 0 and 1.\"\"\"\n    return SequenceMatcher(None, str1, str2).ratio()\n\n# Example\ns1 = \"Acme Corp\"\ns2 = \"Acme Corporation\"\nprint(f\"Similarity: {get_similarity(s1, s2)}\")\n```\n\n#### Finding Best Match in a List\n\n```python\nfrom difflib import get_close_matches\n\nword = \"appel\"\npossibilities = [\"ape\", \"apple\", \"peach\", \"puppy\"]\nmatches = get_close_matches(word, possibilities, n=1, cutoff=0.6)\nprint(matches)\n# Output: ['apple']\n```\n\n### rapidfuzz (Recommended for Performance)\n\nIf `rapidfuzz` is available (pip install rapidfuzz), it is much faster and offers more metrics.\n\n```python\nfrom rapidfuzz import fuzz, process\n\n# Simple Ratio\nscore = fuzz.ratio(\"this is a test\", \"this is a test!\")\nprint(score)\n\n# Partial Ratio (good for substrings)\nscore = fuzz.partial_ratio(\"this is a test\", \"this is a test!\")\nprint(score)\n\n# Extraction\nchoices = [\"Atlanta Falcons\", \"New York Jets\", \"New York Giants\", \"Dallas Cowboys\"]\nbest_match = process.extractOne(\"new york jets\", choices)\nprint(best_match)\n# Output: ('New York Jets', 100.0, 1)\n```\n\n## Common Patterns\n\n### Normalization before Matching\n\nAlways normalize strings before comparing to improve accuracy.\n\n```python\nimport re\n\ndef normalize(text):\n    # Convert to lowercase\n    text = text.lower()\n    # Remove special characters\n    text = re.sub(r'[^\\w\\s]', '', text)\n    # Normalize whitespace\n    text = \" \".join(text.split())\n    # Common abbreviations\n    text = text.replace(\"limited\", \"ltd\").replace(\"corporation\", \"corp\")\n    return text\n\ns1 = \"Acme  Corporation, Inc.\"\ns2 = \"acme corp inc\"\nprint(normalize(s1) == normalize(s2))\n```\n\n### Entity Resolution\n\nWhen matching a list of dirty names to a clean database:\n\n```python\nclean_names = [\"Google LLC\", \"Microsoft Corp\", \"Apple Inc\"]\ndirty_names = [\"google\", \"Microsft\", \"Apple\"]\n\nresults = {}\nfor dirty in dirty_names:\n    # simple containment check first\n    match = None\n    for clean in clean_names:\n        if dirty.lower() in clean.lower():\n            match = clean\n            break\n\n    # fallback to fuzzy\n    if not match:\n        matches = get_close_matches(dirty, clean_names, n=1, cutoff=0.6)\n        if matches:\n            match = matches[0]\n\n    results[dirty] = match\n```\n\n## Reconciliation Output Hygiene (IDs, nulls, and fraud-style rules)\n\nFuzzy matching is often one step in a larger reconciliation pipeline (e.g., matching an invoice vendor name to an approved vendor list, then validating PO\u002Famount\u002FIBAN).\n\n### Treat identifiers as data with explicit validity\n\nWhen an identifier (PO number, vendor id, etc.) is **missing or invalid under the task rules**, represent it as **null** in JSON outputs rather than propagating an unrecognized string.\n\nWhy: downstream evaluators and consumers typically interpret `null` as “not present \u002F not valid”, whereas a string like `\"PO-INVALID\"` looks like a real PO value. In the representative run, a report entry used reason `\"Invalid PO\"` but kept `po_number: \"PO-INVALID\"`, and the verifier expected `po_number: null`.\n\n### Make reason-driven field setting explicit\n\nIf your pipeline chooses a fraud `reason` from an ordered set of criteria, apply consistent field conventions:\n\n- If the selected reason implies the PO is not usable (e.g., **Invalid PO**), set `po_number` to `null` even if the raw invoice text contained a token.\n- If the selected reason is **Unknown Vendor**, still record the raw vendor name string for transparency, but ensure any dependent IDs that cannot be validated are `null` unless the task explicitly asks otherwise.\n\n### Minimal pattern for safe structured output\n\n```python\n# pseudo-code\nif reason == \"Invalid PO\":\n    out_po_number = None\nelif po_number_present:\n    out_po_number = po_number\nelse:\n    out_po_number = None\n\nrecord = {\n  \"po_number\": out_po_number,\n  # ... other fields\n}\n```\n\nThis is not task-instance specific; it is a general reconciliation best practice: **don’t echo invalid identifiers as if they were validated keys**.\n","# Fuzzy Matching Guide\n\n## Overview\n\nThis skill provides methods to compare strings and find the best matches using Levenshtein distance and other similarity metrics. It is essential when joining datasets on string keys that are not identical.\n\n## Quick Start\n\n```python\nfrom difflib import SequenceMatcher\n\ndef similarity(a, b):\n    return SequenceMatcher(None, a, b).ratio()\n\nprint(similarity(\"Apple Inc.\", \"Apple Incorporated\"))\n# Output: 0.7...\n```\n\n## Python Libraries\n\n### difflib (Standard Library)\n\nThe `difflib` module provides classes and functions for comparing sequences.\n\n#### Basic Similarity\n\n```python\nfrom difflib import SequenceMatcher\n\ndef get_similarity(str1, str2):\n    \"\"\"Returns a ratio between 0 and 1.\"\"\"\n    return SequenceMatcher(None, str1, str2).ratio()\n\n# Example\ns1 = \"Acme Corp\"\ns2 = \"Acme Corporation\"\nprint(f\"Similarity: {get_similarity(s1, s2)}\")\n```\n\n#### Finding Best Match in a List\n\n```python\nfrom difflib import get_close_matches\n\nword = \"appel\"\npossibilities = [\"ape\", \"apple\", \"peach\", \"puppy\"]\nmatches = get_close_matches(word, possibilities, n=1, cutoff=0.6)\nprint(matches)\n# Output: ['apple']\n```\n\n### rapidfuzz (Recommended for Performance)\n\nIf `rapidfuzz` is available (pip install rapidfuzz), it is much faster and offers more metrics.\n\n```python\nfrom rapidfuzz import fuzz, process\n\n# Simple Ratio\nscore = fuzz.ratio(\"this is a test\", \"this is a test!\")\nprint(score)\n\n# Partial Ratio (good for substrings)\nscore = fuzz.partial_ratio(\"this is a test\", \"this is a test!\")\nprint(score)\n\n# Extraction\nchoices = [\"Atlanta Falcons\", \"New York Jets\", \"New York Giants\", \"Dallas Cowboys\"]\nbest_match = process.extractOne(\"new york jets\", choices)\nprint(best_match)\n# Output: ('New York Jets', 100.0, 1)\n```\n\n## Common Patterns\n\n### Normalization before Matching\n\nAlways normalize strings before comparing to improve accuracy.\n\n```python\nimport re\n\ndef normalize(text):\n    # Convert to lowercase\n    text = text.lower()\n    # Remove special characters\n    text = re.sub(r'[^\\w\\s]', '', text)\n    # Normalize whitespace\n    text = \" \".join(text.split())\n    # Common abbreviations\n    text = text.replace(\"limited\", \"ltd\").replace(\"corporation\", \"corp\")\n    return text\n\ns1 = \"Acme  Corporation, Inc.\"\ns2 = \"acme corp inc\"\nprint(normalize(s1) == normalize(s2))\n```\n\n### Entity Resolution\n\nWhen matching a list of dirty names to a clean database:\n\n```python\nclean_names = [\"Google LLC\", \"Microsoft Corp\", \"Apple Inc\"]\ndirty_names = [\"google\", \"Microsft\", \"Apple\"]\n\nresults = {}\nfor dirty in dirty_names:\n    # simple containment check first\n    match = None\n    for clean in clean_names:\n        if dirty.lower() in clean.lower():\n            match = clean\n            break\n\n    # fallback to fuzzy\n    if not match:\n        matches = get_close_matches(dirty, clean_names, n=1, cutoff=0.6)\n        if matches:\n            match = matches[0]\n\n    results[dirty] = match\n```\n\n## Reconciliation Output Hygiene (IDs, nulls, and fraud-style rules)\n\nFuzzy matching is often one step in a larger reconciliation pipeline (e.g., matching an invoice vendor name to an approved vendor list, then validating PO\u002Famount\u002FIBAN).\n\n### Treat identifiers as data with explicit validity\n\nWhen an identifier (PO number, vendor id, etc.) is **missing or invalid under the task rules**, represent it as **null** in JSON outputs rather than propagating an unrecognized string.\n\nWhy: downstream evaluators and consumers typically interpret `null` as “not present \u002F not valid”, whereas a string like `\"PO-INVALID\"` looks like a real PO value. In the representative run, a report entry used reason `\"Invalid PO\"` but kept `po_number: \"PO-INVALID\"`, and the verifier expected `po_number: null`.\n\n### Make reason-driven field setting explicit\n\nIf your pipeline chooses a fraud `reason` from an ordered set of criteria, apply consistent field conventions:\n\n- If the selected reason implies the PO is not usable (e.g., **Invalid PO**), set `po_number` to `null` even if the raw invoice text contained a token.\n- If the selected reason is **Unknown Vendor**, still record the raw vendor name string for transparency, but ensure any dependent IDs that cannot be validated are `null` unless the task explicitly asks otherwise.\n\n### Minimal pattern for safe structured output\n\n```python\n# pseudo-code\nif reason == \"Invalid PO\":\n    out_po_number = None\nelif po_number_present:\n    out_po_number = po_number\nelse:\n    out_po_number = None\n\nrecord = {\n  \"po_number\": out_po_number,\n  # ... other fields\n}\n```\n\nThis is not task-instance specific; it is a general reconciliation best practice: **don’t echo invalid identifiers as if they were validated keys**.\n",{"name":5,"license":15,"description":7},"main","cdcb6d230f3b15c3c33f8a2bdd16439f43602cbe","evaluation\u002Fresults\u002FSkillAxe\u002Ftasks\u002Finvoice-fraud-detection\u002Frepaired_run\u002Finputs\u002Fskills\u002Ffuzzy-match",1,"https:\u002F\u002Fgithub.com\u002Flinyouxue\u002Fskill-repair-benchmark","https:\u002F\u002Fgithub.com\u002Flinyouxue","2026-11-14 20:28:35",{"source":8,"skillName":5,"commands":32,"prompt":42,"zipAvailable":43},[33,36,39],{"tool":34,"command":35},"npx","npx skills add linyouxue\u002Fskill-repair-benchmark --skill fuzzy-match",{"tool":37,"command":38},"bunx","bunx skills add linyouxue\u002Fskill-repair-benchmark --skill fuzzy-match",{"tool":40,"command":41},"pnpm","pnpm dlx skills add linyouxue\u002Fskill-repair-benchmark --skill fuzzy-match","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Flinyouxue\u002Fskill-repair-benchmark\u002Fblob\u002Fmain\u002Fevaluation\u002Fresults\u002FSkillAxe\u002Ftasks\u002Finvoice-fraud-detection\u002Frepaired_run\u002Finputs\u002Fskills\u002Ffuzzy-match\u002FSKILL.md",true,[],[46,54,61,68,73,78],{"skillNo":47,"name":48,"nameZh":6,"description":49,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":50,"licenseFileUrl":6,"sourceHtmlUrl":51,"categories":52,"keywords":53,"publishedAt":19},"01M7E1HGVRY73NBJ022CJG13X6","video-frame-extraction","Extract frames from video files and save them as images using OpenCV","Apache-2.0","https:\u002F\u002Fgithub.com\u002Flinyouxue\u002Fskill-repair-benchmark\u002Fblob\u002Fmain\u002Fjpg-ocr-stat\u002Fresult\u002Fskilltta-claude4.7\u002Fgroup_03\u002Frun_03\u002Finputs\u002Fskills\u002Fvideo-frame-extraction\u002FSKILL.md",[],[],{"skillNo":55,"name":56,"nameZh":6,"description":57,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":50,"licenseFileUrl":6,"sourceHtmlUrl":58,"categories":59,"keywords":60,"publishedAt":19},"01M7E1HAG5R7CJ9C8NYJS771MB","openai-vision","Analyze images and multi-frame sequences using OpenAI GPT vision models","https:\u002F\u002Fgithub.com\u002Flinyouxue\u002Fskill-repair-benchmark\u002Fblob\u002Fmain\u002Fjpg-ocr-stat\u002Fresult\u002Fskilltta-claude4.7\u002Fgroup_03\u002Frun_03\u002Finputs\u002Fskills\u002Fopenai-vision\u002FSKILL.md",[],[],{"skillNo":62,"name":63,"nameZh":6,"description":64,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":50,"licenseFileUrl":6,"sourceHtmlUrl":65,"categories":66,"keywords":67,"publishedAt":19},"01M7E1H7DADKB6QBWKC2KQ63SZ","image-ocr","Extract text content from images using Tesseract OCR via Python, including structured field extraction from receipts and forms","https:\u002F\u002Fgithub.com\u002Flinyouxue\u002Fskill-repair-benchmark\u002Fblob\u002Fmain\u002Fjpg-ocr-stat\u002Fresult\u002Fskilltta-claude4.7\u002Fgroup_03\u002Frun_03\u002Finputs\u002Fskills\u002Fimage-ocr\u002FSKILL.md",[],[],{"skillNo":69,"name":48,"nameZh":6,"description":49,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":50,"licenseFileUrl":6,"sourceHtmlUrl":70,"categories":71,"keywords":72,"publishedAt":19},"01M7E1H0WBNT43TENGQY7AVZHD","https:\u002F\u002Fgithub.com\u002Flinyouxue\u002Fskill-repair-benchmark\u002Fblob\u002Fmain\u002Fjpg-ocr-stat\u002Fresult\u002Fskilltta-claude4.7\u002Fgroup_03\u002Frun_02\u002Finputs\u002Fskills\u002Fvideo-frame-extraction\u002FSKILL.md",[],[],{"skillNo":74,"name":56,"nameZh":6,"description":57,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":50,"licenseFileUrl":6,"sourceHtmlUrl":75,"categories":76,"keywords":77,"publishedAt":19},"01M7E1GTC7D4N4Q2EF03875J9E","https:\u002F\u002Fgithub.com\u002Flinyouxue\u002Fskill-repair-benchmark\u002Fblob\u002Fmain\u002Fjpg-ocr-stat\u002Fresult\u002Fskilltta-claude4.7\u002Fgroup_03\u002Frun_02\u002Finputs\u002Fskills\u002Fopenai-vision\u002FSKILL.md",[],[],{"skillNo":79,"name":63,"nameZh":6,"description":64,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":50,"licenseFileUrl":6,"sourceHtmlUrl":80,"categories":81,"keywords":82,"publishedAt":19},"01M7E1GQ92CWFMGH53KSSE4WER","https:\u002F\u002Fgithub.com\u002Flinyouxue\u002Fskill-repair-benchmark\u002Fblob\u002Fmain\u002Fjpg-ocr-stat\u002Fresult\u002Fskilltta-claude4.7\u002Fgroup_03\u002Frun_02\u002Finputs\u002Fskills\u002Fimage-ocr\u002FSKILL.md",[],[]]