ggml-org · hanishkvc · Oct 11, 2025 · Oct 11, 2025 · Oct 11, 2025 · Oct 11, 2025
diff --git a/tools/server/public_simplechat/datautils.mjs b/tools/server/public_simplechat/datautils.mjs
@@ -125,6 +125,9 @@ export function trim_hist_garbage_at_end(sIn, maxType, maxUniq, maxMatchLenThres
     let iNum = 0;
     let iOth = 0;
     // Learn
+    /**
+     * @type {Object<string, number>}
+     */
     let hist = {};
     let iUniq = 0;
     for(let i=0; i<maxMatchLenThreshold; i++) {

diff --git a/tools/server/public_simplechat/index.html b/tools/server/public_simplechat/index.html
@@ -11,20 +11,24 @@
         <script type="importmap">
             {
                 "imports": {
+                    "simplechat": "./simplechat.js",
                     "datautils": "./datautils.mjs",
                     "ui": "./ui.mjs"
                 }
             }
         </script>
-        <script src="simplechat.js" type="module" defer></script>
+        <script src="main.js" type="module" defer></script>
         <link rel="stylesheet" href="simplechat.css" />
     </head>
     <body>
         <div class="samecolumn" id="fullbody">
 
             <div class="sameline" id="heading">
                 <p class="heading flex-grow" > <b> SimpleChat </b> </p>
-                <button id="settings">Settings</button>
+                <div>
+                <button id="clearchat">&#x1F9F9; ClearChat</button>
+                <button id="settings">&#x2699; Settings</button>
+                </div>
             </div>
 
             <div id="sessions-div" class="sameline"></div>
@@ -40,6 +44,19 @@
                 <p> You need to have javascript enabled.</p>
             </div>
 
+            <hr>
+            <div id="tool-div">
+                <div>
+                <div class="sameline">
+                    <textarea id="toolname-in" class="flex-grow" rows="1" placeholder="name of tool to run"></textarea>
+                    <button id="tool-btn">run tool</button>
+                </div>
+                </div>
+                <div class="sameline">
+                <textarea id="toolargs-in" class="flex-grow" rows="2" placeholder="arguments to pass to the specified tool"></textarea>
+                </div>
+            </div>
+
             <hr>
             <div class="sameline">
                 <textarea id="user-in" class="flex-grow" rows="2" placeholder="enter your query to the ai model here" ></textarea>

diff --git a/tools/server/public_simplechat/local.tools/debug.py b/tools/server/public_simplechat/local.tools/debug.py
@@ -0,0 +1,24 @@
+# Helpers for debugging
+# by Humans for All
+
+
+import time
+
+gMe = { '--debug' : False }
+
+
+def setup(bEnable):
+    global gMe
+    gMe['--debug'] = bEnable
+
+
+def dump(meta: dict, data: dict):
+    if not gMe['--debug']:
+        return
+    timeTag = f"{time.time():0.12f}"
+    with open(f"/tmp/simpleproxy.{timeTag}.meta", '+w') as f:
+        for k in meta:
+            f.write(f"\n\n\n\n{k}:{meta[k]}\n\n\n\n")
+    with open(f"/tmp/simpleproxy.{timeTag}.data", '+w') as f:
+        for k in data:
+            f.write(f"\n\n\n\n{k}:{data[k]}\n\n\n\n")
diff --git a/tools/server/public_simplechat/local.tools/filemagic.py b/tools/server/public_simplechat/local.tools/filemagic.py
@@ -0,0 +1,81 @@
+# Handle file related helpers, be it a local file or one on the internet
+# by Humans for All
+
+import urllib.request
+import urllib.parse
+import debug
+from dataclasses import dataclass
+
+
+@dataclass(frozen=True)
+class Response:
+    """
+    Used to return result wrt urlreq helper below.
+    """
+    callOk: bool
+    statusCode: int
+    statusMsg: str = ""
+    contentType: str = ""
+    contentData: bytes = b""
+
+
+
+def get_from_web(url: str, tag: str, inContentType: str, inHeaders: dict[str, str|None]):
+    """
+    Get the url specified from web.
+
+    If passed header doesnt contain certain useful http header entries,
+    some predefined defaults will be used in place.
+    """
+    try:
+        hUA = inHeaders.get('User-Agent', None)
+        if not hUA:
+            hUA = 'Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:132.0) Gecko/20100101 Firefox/132.0'
+        hAL = inHeaders.get('Accept-Language', None)
+        if not hAL:
+            hAL = "en-US,en;q=0.9"
+        hA = inHeaders.get('Accept', None)
+        if not hA:
+            hA = "text/html,*/*"
+        headers = {
+            'User-Agent': hUA,
+            'Accept': hA,
+            'Accept-Language': hAL
+        }
+        req = urllib.request.Request(url, headers=headers)
+        # Get requested url
+        print(f"DBUG:{tag}:Req:{req.full_url}:{req.headers}")
+        with urllib.request.urlopen(req, timeout=10) as response:
+            contentData = response.read()
+            statusCode = response.status or 200
+            statusMsg = response.msg or ""
+            contentType = response.getheader('Content-Type') or inContentType
+            print(f"DBUG:FM:GFW:Resp:{response.status}:{response.msg}")
+            debug.dump({ 'url': req.full_url, 'headers': req.headers, 'ctype': contentType }, { 'cdata': contentData })
+        return Response(True, statusCode, statusMsg, contentType, contentData)
+    except Exception as exc:
+        return Response(False, 502, f"WARN:{tag}:Failed:{exc}")
+
+
+def get_from_local(urlParts: urllib.parse.ParseResult, tag: str, inContentType: str):
+    """
+    Get the requested file from the local filesystem
+    """
+    try:
+        fPdf = open(urlParts.path, 'rb')
+        dPdf = fPdf.read()
+        return Response(True, 200, "", inContentType, dPdf)
+    except Exception as exc:
+        return Response(False, 502, f"WARN:{tag}:Failed:{exc}")
+
+
+def get_file(url: str, tag: str, inContentType: str, inHeaders: dict[str, str|None]={}):
+    """
+    Based on the scheme specified in the passed url,
+    either get from local file system or from the web.
+    """
+    urlParts = urllib.parse.urlparse(url)
+    if urlParts.scheme == "file":
+        return get_from_local(urlParts, tag, inContentType)
+    else:
+        return get_from_web(url, tag, inContentType, inHeaders)
diff --git a/tools/server/public_simplechat/local.tools/pdfmagic.py b/tools/server/public_simplechat/local.tools/pdfmagic.py
@@ -0,0 +1,94 @@
+# Helper to manage pdf related requests
+# by Humans for All
+
+import urllib.parse
+import urlvalidator as uv
+import filemagic as mFile
+from typing import TYPE_CHECKING, Any
+
+if TYPE_CHECKING:
+    from simpleproxy import ProxyHandler
+
+
+PDFOUTLINE_MAXDEPTH=4
+
+
+def extract_pdfoutline(ol: Any, prefix: list[int]):
+    """
+    Helps extract the pdf outline recursively, along with its numbering.
+    """
+    if (len(prefix) > PDFOUTLINE_MAXDEPTH):
+        return ""
+    if type(ol).__name__ != type([]).__name__:
+        prefix[-1] += 1
+        return f"{".".join(map(str,prefix))}:{ol['/Title']}\n"
+    olText = ""
+    prefix.append(0)
+    for (i,iol) in enumerate(ol):
+        olText += extract_pdfoutline(iol, prefix)
+    prefix.pop()
+    return olText
+
+
+def process_pdftext(url: str, startPN: int, endPN: int):
+    """
+    Extract textual content from given pdf.
+
+    * Validate the got url.
+    * Get the pdf file.
+    * Extract textual contents of the pdf from given start page number to end page number (inclusive).
+        * if -1 | 0 is specified wrt startPN, the actual starting page number (rather 1) will be used.
+        * if -1 | 0 is specified wrt endPN, the actual ending page number will be used.
+
+    NOTE: Page numbers start from 1, while the underlying list data structure index starts from 0
+    """
+    import pypdf
+    import io
+    gotVU = uv.validate_url(url, "HandlePdfText")
+    if not gotVU.callOk:
+        return { 'status': gotVU.statusCode, 'msg': gotVU.statusMsg }
+    gotFile = mFile.get_file(url, "ProcessPdfText", "application/pdf", {})
+    if not gotFile.callOk:
+        return { 'status': gotFile.statusCode, 'msg': gotFile.statusMsg, 'data': gotFile.contentData}
+    tPdf = ""
+    oPdf = pypdf.PdfReader(io.BytesIO(gotFile.contentData))
+    if (startPN <= 0):
+        startPN = 1
+    if (endPN <= 0) or (endPN > len(oPdf.pages)):
+        endPN = len(oPdf.pages)
+    # Add the pdf outline, if available
+    outlineGot = extract_pdfoutline(oPdf.outline, [])
+    if outlineGot:
+        tPdf += f"\n\nOutline Start\n\n{outlineGot}\n\nOutline End\n\n"
+    # Add the pdf page contents
+    for i in range(startPN, endPN+1):
+        pd = oPdf.pages[i-1]
+        tPdf = tPdf + pd.extract_text()
+    return { 'status': 200, 'msg': "PdfText Response follows", 'data': tPdf }
+
+
+def handle_pdftext(ph: 'ProxyHandler', pr: urllib.parse.ParseResult):
+    """
+    Handle requests to pdftext path, which is used to extract plain text
+    from the specified pdf file.
+    """
+    queryParams = urllib.parse.parse_qs(pr.query)
+    url = queryParams['url'][0]
+    startP = queryParams.get('startPageNumber', -1)
+    if isinstance(startP, list):
+        startP = int(startP[0])
+    endP = queryParams.get('endPageNumber', -1)
+    if isinstance(endP, list):
+        endP = int(endP[0])
+    print(f"INFO:HandlePdfText:Processing:{url}:{startP}:{endP}...")
+    gotP2T = process_pdftext(url, startP, endP)
+    if (gotP2T['status'] != 200):
+        ph.send_error(gotP2T['status'], gotP2T['msg'] )
+        return
+    ph.send_response(gotP2T['status'], gotP2T['msg'])
+    ph.send_header('Content-Type', 'text/text')
+    # Add CORS for browser fetch, just in case
+    ph.send_header('Access-Control-Allow-Origin', '*')
+    ph.end_headers()
+    print(f"INFO:HandlePdfText:ExtractedText:{url}...")
+    ph.wfile.write(gotP2T['data'].encode('utf-8'))
diff --git a/tools/server/public_simplechat/local.tools/simpleproxy.json b/tools/server/public_simplechat/local.tools/simpleproxy.json
@@ -0,0 +1,55 @@
+{
+    "allowed.schemes": [
+        "file",
+        "http",
+        "https"
+    ],
+    "allowed.domains": [
+        ".*\\.wikipedia\\.org$",
+        ".*\\.bing\\.com$",
+        "^www\\.bing\\.com$",
+        ".*\\.yahoo\\.com$",
+        "^search\\.yahoo\\.com$",
+        ".*\\.brave\\.com$",
+        "^search\\.brave\\.com$",
+        "^brave\\.com$",
+        ".*\\.duckduckgo\\.com$",
+        "^duckduckgo\\.com$",
+        ".*\\.google\\.com$",
+        "^google\\.com$",
+        ".*\\.arxiv\\.org$",
+        "^arxiv\\.org$",
+        ".*\\.nature\\.com$",
+        ".*\\.science\\.org$",
+        "^apnews\\.com$",
+        ".*\\.apnews\\.com$",
+        ".*\\.reuters\\.com$",
+        ".*\\.bloomberg\\.com$",
+        ".*\\.forbes\\.com$",
+        ".*\\.npr\\.org$",
+        ".*\\.cnn\\.com$",
+        ".*\\.theguardian\\.com$",
+        ".*\\.bbc\\.com$",
+        ".*\\.france24\\.com$",
+        ".*\\.dw\\.com$",
+        ".*\\.jpost\\.com$",
+        ".*\\.aljazeera\\.com$",
+        ".*\\.alarabiya\\.net$",
+        ".*\\.rt\\.com$",
+        "^tass\\.com$",
+        ".*\\.channelnewsasia\\.com$",
+        ".*\\.scmp\\.com$",
+        ".*\\.nikkei\\.com$",
+        ".*\\.nhk\\.or\\.jp$",
+        ".*\\.indiatoday\\.in$",
+        "^theprint\\.in$",
+        ".*\\.ndtv\\.com$",
+        "^lwn\\.net$",
+        "^arstechnica\\.com$",
+        ".*\\.linkedin\\.com$",
+        ".*\\.github\\.io$",
+        "^github\\.com$",
+        ".*\\.github\\.com$"
+    ],
+    "bearer.insecure": "NeverSecure"
+}