refactor: transition to Claude Plugin Marketplace with Native MCP and shell scripts

plugin/skills/targeting/SKILL.md

@@ -81,15 +81,15 @@ A search result is considered **"High Noise"** if **8 or more** of the top 20 sn
 
 1. **Run Merge Command**:
    - Execute the following command to combine the CSV files and remove duplicates.
-   - **Important**: Use `patent-kit` command, NOT `MCP tool`.
-   - Command: `patent-kit merge --input-dir 1-targeting/csv --output 1-targeting/target.jsonl`
+   - **Important**: Use `./plugin/skills/targeting/scripts/shell/merge.sh` (Mac/Linux) or `.\plugin\skills\targeting\scripts\powershell\merge.ps1` (Windows), NOT `MCP tool`.
+   - Command: `./plugin/skills/targeting/scripts/shell/merge.sh 1-targeting/csv 1-targeting/target.jsonl`
 
 2. **Verify Output**:
    - Check that `1-targeting/target.jsonl` has been created.
    - This file contains the consolidated list of unique patents to be screened/evaluated.
 
 3. **Check Count**:
-   - The `patent-kit merge` command output displays the number of unique patents (e.g., `Merged 150 unique patents...`).
+   - The merge command output displays the number of unique patents (e.g., `Merged 150 unique patents...`).
    - Confirm this count to understand the volume of patents to be screened.
 
 ### Output
@@ -139,6 +139,6 @@ Actions:
 
 # Troubleshooting
 
-Error: "patent-kit command not found"
-Cause: Python script path is incorrect or environment is not fully setup.
-Solution: Ensure plugin/skills/targeting/scripts/merge.py is executable and Python 3 is installed.
+Error: "Permission denied" when running merge.sh
+Cause: The script lacks execution permissions.
+Solution: Run `chmod +x plugin/skills/targeting/scripts/shell/merge.sh`.

plugin/skills/targeting/scripts/powershell/merge.ps1

@@ -1,9 +1,74 @@
-param (
-    [string]$InputPath = "1-targeting/csv",
-    [string]$OutputPath = "1-targeting/target.jsonl"
+param(
+    [string]$InputDir = "1-targeting\csv",
+    [string]$OutputFile = "1-targeting\target.jsonl"
 )
 
-$ScriptDir = Split-Path -Parent $MyInvocation.MyCommand.Definition
-$MergePy = Join-Path $ScriptDir "..\merge.py"
+if (-not (Test-Path $InputDir)) {
+    Write-Host "Directory not found: $InputDir. Please create it and place CSV files there."
+    exit
+}
 
-python3 $MergePy $InputPath $OutputPath
+$UniquePatents = @{}
+$FileCount = 0
+
+Get-ChildItem -Path $InputDir -Filter "*.csv" | ForEach-Object {
+    $FilePath = $_.FullName
+    Write-Host "Processing $($_.Name)"
+    $FileCount++
+
+    $Lines = Get-Content -Path $FilePath -Encoding UTF8
+
+    $CsvStartIndex = 0
+    for ($i = 0; $i -lt $Lines.Count; $i++) {
+        $Line = $Lines[$i].Trim()
+        if ($Line -ne "" -and -not $Line.StartsWith("search URL:")) {
+            $CsvStartIndex = $i
+            break
+        }
+    }
+
+    if ($CsvStartIndex -lt $Lines.Count) {
+        $CsvData = $Lines[$CsvStartIndex..($Lines.Count - 1)] | ConvertFrom-Csv
+
+        foreach ($Row in $CsvData) {
+            if (-not [string]::IsNullOrWhiteSpace($Row.id)) {
+                $UniquePatents[$Row.id] = $Row
+            }
+        }
+    }
+}
+
+if ($FileCount -eq 0) {
+    Write-Host "No CSV files found in $InputDir"
+    exit
+}
+
+$OutputDir = Split-Path $OutputFile -Parent
+if ($OutputDir -and -not (Test-Path $OutputDir)) {
+    New-Item -ItemType Directory -Force -Path $OutputDir | Out-Null
+}
+
+$OutFileStream = [System.IO.StreamWriter]::new((Resolve-Path -Path ".").ProviderPath + "\" + $OutputFile, $false, [System.Text.Encoding]::UTF8)
+
+foreach ($Record in $UniquePatents.Values) {
+    $FilteredRecord = [ordered]@{}
+    foreach ($Prop in $Record.PSObject.Properties) {
+        $K = $Prop.Name
+        $V = $Prop.Value
+        if (-not [string]::IsNullOrEmpty($K) -and $K -notmatch 'link' -and $K -ne 'inventor/author') {
+            $FilteredRecord[$K] = $V
+        }
+    }
+
+    if ($FilteredRecord.Contains('id')) {
+        $FilteredRecord['id'] = $FilteredRecord['id'] -replace '-', ''
+    }
+
+    $Json = $FilteredRecord | ConvertTo-Json -Depth 5 -Compress
+    $OutFileStream.WriteLine($Json)
+}
+
+$OutFileStream.Close()
+
+$PatentCount = $UniquePatents.Count
+Write-Host "Merged $PatentCount unique patents from $FileCount files into $OutputFile"

plugin/skills/targeting/scripts/shell/merge.sh

@@ -1,5 +1,130 @@
 #!/bin/bash
-set -e
+INPUT_DIR="${1:-1-targeting/csv}"
+OUTPUT_FILE="${2:-1-targeting/target.jsonl}"
 
-DIR="$( cd "$( dirname "${BASH_SOURCE[0]}" )" &> /dev/null && pwd )"
-python3 "$DIR/../merge.py" "$@"
+if [ ! -d "$INPUT_DIR" ]; then
+    echo "Directory not found: $INPUT_DIR. Please create it and place CSV files there."
+    exit 0
+fi
+
+mkdir -p "$(dirname "$OUTPUT_FILE")"
+
+# Use Node.js to safely parse CSV and output JSONL
+node -e "
+const fs = require('fs');
+const path = require('path');
+
+const inputDir = process.argv[1];
+const outputFile = process.argv[2];
+
+let uniquePatents = {};
+let fileCount = 0;
+
+const files = fs.readdirSync(inputDir).filter(f => f.endsWith('.csv'));
+
+function parseCSV(csvText) {
+    const lines = [];
+    let currentLine = [];
+    let currentCell = '';
+    let inQuotes = false;
+
+    for (let i = 0; i < csvText.length; i++) {
+        const char = csvText[i];
+        const nextChar = csvText[i+1];
+
+        if (inQuotes) {
+            if (char === '\\\"' && nextChar === '\\\"') {
+                currentCell += '\\\"';
+                i++;
+            } else if (char === '\\\"') {
+                inQuotes = false;
+            } else {
+                currentCell += char;
+            }
+        } else {
+            if (char === '\\\"') {
+                inQuotes = true;
+            } else if (char === ',') {
+                currentLine.push(currentCell);
+                currentCell = '';
+            } else if (char === '\\n' || char === '\\r') {
+                if (char === '\\r' && nextChar === '\\n') i++;
+                currentLine.push(currentCell);
+                lines.push(currentLine);
+                currentLine = [];
+                currentCell = '';
+            } else {
+                currentCell += char;
+            }
+        }
+    }
+    if (currentCell !== '' || currentLine.length > 0) {
+        currentLine.push(currentCell);
+        lines.push(currentLine);
+    }
+    return lines;
+}
+
+for (const file of files) {
+    const filePath = path.join(inputDir, file);
+    console.log('Processing ' + file);
+    fileCount++;
+
+    const content = fs.readFileSync(filePath, 'utf8');
+    const textLines = content.split(/\r?\n/);
+
+    let csvStartIndex = 0;
+    for (let i = 0; i < textLines.length; i++) {
+        const line = textLines[i].trim();
+        if (line && !line.startsWith('search URL:')) {
+            csvStartIndex = i;
+            break;
+        }
+    }
+
+    const csvContent = textLines.slice(csvStartIndex).join('\n');
+    const parsed = parseCSV(csvContent);
+    if (parsed.length < 2) continue;
+
+    const headers = parsed[0];
+
+    for (let i = 1; i < parsed.length; i++) {
+        const row = parsed[i];
+        if (row.length === 0 || (row.length === 1 && !row[0])) continue;
+
+        const obj = {};
+        for (let j = 0; j < headers.length; j++) {
+            obj[headers[j]] = row[j] || '';
+        }
+
+        if (obj['id']) {
+            uniquePatents[obj['id']] = obj;
+        }
+    }
+}
+
+if (fileCount === 0) {
+    console.log('No CSV files found in ' + inputDir);
+    process.exit(0);
+}
+
+const outData = [];
+for (const key in uniquePatents) {
+    const record = uniquePatents[key];
+    const filteredRecord = {};
+
+    for (const k in record) {
+        if (k && !k.includes('link') && k !== 'inventor/author') {
+            filteredRecord[k] = record[k];
+        }
+    }
+
+    if (filteredRecord['id']) {
+        filteredRecord['id'] = filteredRecord['id'].replace(/-/g, '');
+    }
+    outData.push(JSON.stringify(filteredRecord));
+}
+
+fs.writeFileSync(outputFile, outData.join('\\n') + '\\n', 'utf8');
+console.log('Merged ' + Object.keys(uniquePatents).length + ' unique patents from ' + fileCount + ' files into ' + outputFile);
+" "$INPUT_DIR" "$OUTPUT_FILE"