toolslm 0.3.9__py3-none-any.whl → 0.3.17__py3-none-any.whl

toolslm/__init__.py

@@ -1 +1 @@
-__version__ = "0.3.9"
+__version__ = "0.3.17"

toolslm/_modidx.py

@@ -49,5 +49,10 @@ d = { 'settings': { 'branch': 'main',
                              'toolslm.xml.mk_doc': ('xml.html#mk_doc', 'toolslm/xml.py'),
                              'toolslm.xml.mk_doctype': ('xml.html#mk_doctype', 'toolslm/xml.py'),
                              'toolslm.xml.nb2xml': ('xml.html#nb2xml', 'toolslm/xml.py'),
+                             'toolslm.xml.parse_gh_url': ('xml.html#parse_gh_url', 'toolslm/xml.py'),
                              'toolslm.xml.read_file': ('xml.html#read_file', 'toolslm/xml.py'),
-                             'toolslm.xml.repo2ctx': ('xml.html#repo2ctx', 'toolslm/xml.py')}}}
+                             'toolslm.xml.repo2ctx': ('xml.html#repo2ctx', 'toolslm/xml.py'),
+                             'toolslm.xml.sym2file': ('xml.html#sym2file', 'toolslm/xml.py'),
+                             'toolslm.xml.sym2folderctx': ('xml.html#sym2folderctx', 'toolslm/xml.py'),
+                             'toolslm.xml.sym2pkgctx': ('xml.html#sym2pkgctx', 'toolslm/xml.py'),
+                             'toolslm.xml.sym2pkgpath': ('xml.html#sym2pkgpath', 'toolslm/xml.py')}}}

toolslm/download.py

@@ -8,6 +8,7 @@ from fastcore.utils import *
 from httpx import get
 from fastcore.meta import delegates
 from urllib.parse import urlparse, urljoin
+from .xml import parse_gh_url
 
 # %% ../03_download.ipynb
 def clean_md(text, rm_comments=True, rm_details=True):

toolslm/funccall.py

@@ -5,7 +5,7 @@ __all__ = ['empty', 'custom_types', 'get_schema', 'python', 'mk_ns', 'call_func'
            'mk_tool']
 
 # %% ../01_funccall.ipynb
-import inspect, json
+import inspect, json, ast
 from collections import abc
 from fastcore.utils import *
 from fastcore.docments import docments
@@ -37,12 +37,20 @@ def _types(t:type)->tuple[str,Optional[str]]:
     else: return tmap.get(t.__name__, "object"), None
 
 # %% ../01_funccall.ipynb
-def _param(name, info):
-    "json schema parameter given `name` and `info` from docments full dict."
+def _param(
+    name, # param name
+    info, # dict from docments
+    evalable=False): # stringify defaults that can't be literal_eval'd?
+    "json schema parameter given `name` and `info` from docments full dict"
     paramt,itemt = _types(info.anno)
     pschema = dict(type=paramt, description=info.docment or "")
     if itemt: pschema["items"] = {"type": itemt}
-    if info.default is not empty: pschema["default"] = info.default
+    if info.default is not empty:
+        if evalable:
+            try: ast.literal_eval(repr(info.default))
+            except: pschema["default"] = str(info.default)
+            else: pschema["default"] = info.default
+        else: pschema["default"] = info.default
     return pschema
 
 # %% ../01_funccall.ipynb
@@ -90,9 +98,9 @@ def _handle_container(origin, args, defs):
     return None
 
 # %% ../01_funccall.ipynb
-def _process_property(name, obj, props, req, defs):
+def _process_property(name, obj, props, req, defs, evalable=False):
     "Process a single property of the schema"
-    p = _param(name, obj)
+    p = _param(name, obj, evalable=evalable)
     props[name] = p
     if obj.default is empty: req[name] = True
 
@@ -103,14 +111,14 @@ def _process_property(name, obj, props, req, defs):
         p.update(_handle_type(obj.anno, defs))
 
 # %% ../01_funccall.ipynb
-def _get_nested_schema(obj):
+def _get_nested_schema(obj, evalable=False, skip_hidden=False):
     "Generate nested JSON schema for a class or function"
     d = docments(obj, full=True)
     props, req, defs = {}, {}, {}
 
     for n, o in d.items():
-        if n != 'return' and n != 'self':
-            _process_property(n, o, props, req, defs)
+        if n != 'return' and n != 'self' and not (skip_hidden and n.startswith('_')):
+            _process_property(n, o, props, req, defs, evalable=evalable)
 
     tkw = {}
     if isinstance(obj, type): tkw['title']=obj.__name__
@@ -120,10 +128,15 @@ def _get_nested_schema(obj):
     return schema
 
 # %% ../01_funccall.ipynb
-def get_schema(f:Union[callable,dict], pname='input_schema')->dict:
+def get_schema(
+    f:Union[callable,dict], # Function to get schema for
+    pname='input_schema',   # Key name for parameters
+    evalable=False,  # stringify defaults that can't be literal_eval'd?
+    skip_hidden=False # skip parameters starting with '_'?
+)->dict: # {'name':..., 'description':..., pname:...}
     "Generate JSON schema for a class, function, or method"
     if isinstance(f, dict): return f
-    schema = _get_nested_schema(f)
+    schema = _get_nested_schema(f, evalable=evalable, skip_hidden=skip_hidden)
     desc = f.__doc__
     assert desc, "Docstring missing!"
     d = docments(f, full=True)

toolslm/xml.py

@@ -2,10 +2,11 @@
 
 # %% auto 0
 __all__ = ['doctype', 'json_to_xml', 'get_mime_text', 'cell2out', 'cell2xml', 'nb2xml', 'mk_doctype', 'mk_doc', 'docs_xml',
-           'read_file', 'files2ctx', 'folder2ctx', 'repo2ctx', 'folder2ctx_cli']
+           'read_file', 'files2ctx', 'folder2ctx', 'sym2file', 'sym2folderctx', 'sym2pkgpath', 'sym2pkgctx',
+           'folder2ctx_cli', 'parse_gh_url', 'repo2ctx']
 
 # %% ../00_xml.ipynb
-import hashlib,xml.etree.ElementTree as ET
+import hashlib, inspect, xml.etree.ElementTree as ET
 from collections import namedtuple
 from ghapi.all import GhApi
 
@@ -49,22 +50,23 @@ def cell2out(o):
     if hasattr(o, 'ename'): return Out(f"{o.ename}: {o.evalue}", type='error')
 
 # %% ../00_xml.ipynb
-def cell2xml(cell, out=True):
+def cell2xml(cell, out=True, ids=True):
     "Convert notebook cell to concise XML format"
     src = ''.join(getattr(cell, 'source', ''))
     f = Code if cell.cell_type=='code' else Md
-    if not out: return f(src)
+    kw = dict(id=cell.id) if ids and hasattr(cell, 'id') else {}
+    if not out: return f(src, **kw)
     parts = [Source(src)]
     out_items = L(getattr(cell,'outputs',[])).map(cell2out).filter()
     if out_items: parts.append(Outs(*out_items))
-    return f(*parts)
+    return f(*parts, **kw)
 
 # %% ../00_xml.ipynb
-def nb2xml(fname=None, nb=None, out=True):
+def nb2xml(fname=None, nb=None, out=True, ids=True):
     "Convert notebook to XML format"
     assert bool(fname)^bool(nb), "Pass either `fname` or `nb`"
     if not nb: nb = dict2obj(fname.read_json())
-    cells_xml = [to_xml(cell2xml(c, out=out), do_escape=False) for c in nb.cells if c.cell_type in ('code','markdown')]
+    cells_xml = [to_xml(cell2xml(c, out=out, ids=ids), do_escape=False) for c in nb.cells if c.cell_type in ('code','markdown')]
     return to_xml(Notebook(*cells_xml), do_escape=False)
 
 # %% ../00_xml.ipynb
@@ -101,7 +103,7 @@ def mk_doc(index:int,  # The document index
 # %% ../00_xml.ipynb
 def docs_xml(docs:list[str],  # The content of each document
              srcs:Optional[list]=None,  # URLs, filenames, etc; each one defaults to `md5(content)` if not provided
-             prefix:bool=True, # Include Anthropic's suggested prose intro?
+             prefix:bool=False, # Include Anthropic's suggested prose intro?
              details:Optional[list]=None, # Optional list of dicts with additional attrs for each doc
              title:str=None # Optional title attr for Documents element
             )->str:
@@ -114,67 +116,82 @@ def docs_xml(docs:list[str],  # The content of each document
     return pre + to_xml(Documents(*docs, **kw), do_escape=False)
 
 # %% ../00_xml.ipynb
-def read_file(fname, out=True, max_size=None):
+def read_file(fname, out=True, max_size=None, ids=True):
     "Read file content, converting notebooks to XML if needed"
     fname = Path(fname)
-    if fname.suffix == '.ipynb': res = nb2xml(fname, out=out)
+    if fname.suffix == '.ipynb': res = nb2xml(fname, out=out, ids=ids)
     else: res = fname.read_text()
     if max_size and len(res)>max_size: return f"[Skipped: {fname.name} exceeds {max_size} bytes]"
     return res
 
 # %% ../00_xml.ipynb
+@delegates(docs_xml)
 def files2ctx(
     fnames:list[Union[str,Path]], # List of file names to add to context
-    prefix:bool=True, # Include Anthropic's suggested prose intro?
     out:bool=True, # Include notebook cell outputs?
     srcs:Optional[list]=None, # Use the labels instead of `fnames`
-    title:str=None, # Optional title attr for Documents element
-    max_size:int=None # Skip files larger than this (bytes)
+    max_size:int=None, # Skip files larger than this (bytes)
+    ids:bool=True,  # Include cell ids in notebooks?
+    **kwargs
 )->str: # XML for LM context
     "Convert files to XML context, handling notebooks"
     fnames = [Path(o) for o in fnames]
-    contents = [read_file(o, out=out, max_size=max_size) for o in fnames]
-    return docs_xml(contents, srcs or fnames, prefix=prefix, title=title)
+    contents = [read_file(o, out=out, max_size=max_size, ids=ids) for o in fnames]
+    return docs_xml(contents, srcs or fnames, **kwargs)
 
 # %% ../00_xml.ipynb
 @delegates(globtastic)
 def folder2ctx(
-    folder:Union[str,Path],
-    prefix:bool=True, # Include Anthropic's suggested prose intro?
+    folder:Union[str,Path], # Folder to read
+    prefix:bool=False, # Include Anthropic's suggested prose intro?
     out:bool=True, # Include notebook cell outputs?
     include_base:bool=True, # Include full path in src?
     title:str=None, # Optional title attr for Documents element
     max_size:int=100_000, # Skip files larger than this (bytes)
+    max_total:int=10_000_000,  # Max total output size in bytes
+    readme_first:bool=False,  # Prioritize README files at start of context?
+    files_only:bool=False,  # Return dict of {filename: size} instead of context?
+    ids:bool=True,  # Include cell ids in notebooks?
     **kwargs
-)->str:
+)->Union[str,dict]:
     "Convert folder contents to XML context, handling notebooks"
     folder = Path(folder)
-    fnames = globtastic(folder, **kwargs)
-    srcs = fnames if include_base else [Path(f).relative_to(folder) for f in fnames]
-    return files2ctx(fnames, prefix=prefix, out=out, srcs=srcs, title=title, max_size=max_size)
+    fnames = pglob(folder, **kwargs)
+    if files_only: return {str(f.relative_to(folder)): f.stat().st_size for f in fnames}
+    if readme_first: fnames = sorted(fnames, key=lambda f: (0 if 'readme' in f.name.lower() else 1, f))
+    srcs = fnames if include_base else [f.relative_to(folder) for f in fnames]
+    res = files2ctx(fnames, prefix=prefix, out=out, srcs=srcs, title=title, max_size=max_size, ids=ids)
+    suf = f"\n\n[TRUNCATED: output size {{_outsz_}} exceeded max size {max_total} bytes]"
+    if max_total and len(res) > max_total: res = truncstr(res, max_total, suf=suf, sizevar='_outsz_')
+    return res
+
+# %% ../00_xml.ipynb
+def sym2file(sym):
+    "Return md string with filepath and contents for a symbol's source file"
+    f = Path(inspect.getfile(sym))
+    return f"- `{f}`\n\n````\n{f.read_text()}\n````"
 
 # %% ../00_xml.ipynb
 @delegates(folder2ctx)
-def repo2ctx(
-    owner:str,  # GitHub repo owner
-    repo:str,   # GitHub repo name
-    ref:str=None,  # Git ref (branch/tag/sha); defaults to repo's default branch
-    **kwargs  # Passed to `folder2ctx`
-)->str:  # XML for LM context
-    "Convert GitHub repo to XML context without cloning"
-    import tempfile, tarfile, io
-    api = GhApi()
-    if ref is None: ref = api.repos.get(owner, repo).default_branch
-    data = api.repos.download_tarball_archive(owner, repo, ref)
-    parts = ' | '.join(f"{k}: {', '.join(v) if isinstance(v, (list,tuple)) else v}"
-        for k,v in kwargs.items() if v)
-    title = f"GitHub repository contents from {owner}/{repo} at ref '{ref}'"
-    if parts: title += f" (filters applied: {parts})"
-    tf = tarfile.open(fileobj=io.BytesIO(data))
-    with tempfile.TemporaryDirectory() as tmp:
-        tf.extractall(tmp, filter='data')
-        subdir = Path(tmp) / tf.getmembers()[0].name.split('/')[0]
-        return folder2ctx(subdir, include_base=False, title=title, **kwargs)
+def sym2folderctx(
+    sym,
+    types:str|list='py',  # list or comma-separated str of ext types from: py, js, java, c, cpp, rb, r, ex, sh, web, doc, cfg
+    skip_file_re=r'^_mod',
+    **kwargs):
+    "Return folder context for a symbol's source file location"
+    return folder2ctx(Path(inspect.getfile(sym)).parent, types=types, skip_file_re=skip_file_re, **kwargs)
+
+# %% ../00_xml.ipynb
+def sym2pkgpath(sym):
+    "Get root package path for a symbol"
+    root = sym.__module__.split('.')[0]
+    return Path(sys.modules[root].__path__[0])
+
+# %% ../00_xml.ipynb
+@delegates(folder2ctx)
+def sym2pkgctx(sym, types:str|list='py', skip_file_re=r'^_mod', **kwargs):
+    "Return repo context for a symbol's root package"
+    return folder2ctx(sym2pkgpath(sym), types=types, skip_file_re=skip_file_re, **kwargs)
 
 # %% ../00_xml.ipynb
 @call_parse
@@ -186,3 +203,44 @@ def folder2ctx_cli(
 )->str: # XML for Claude context
     "CLI to convert folder contents to XML context, handling notebooks"
     print(folder2ctx(folder, out=out, **kwargs))
+
+# %% ../00_xml.ipynb
+def parse_gh_url(url):
+    "Parse GitHub URL into (owner, repo, type, ref, path) or None"
+    m = re.match(r'https?://(?:www\.)?github\.com/([^/]+)/([^/]+)(?:/([^/]+)(?:/([^/]+)(?:/(.+))?)?)?', url)
+    return dict(zip('owner repo typ ref path'.split(), m.groups())) if m else None
+
+# %% ../00_xml.ipynb
+@delegates(folder2ctx)
+def repo2ctx(
+    owner:str,  # GitHub repo owner or "owner/repo" or a full github URL
+    repo:str=None,   # GitHub repo name (leave empty if using "owner/repo" or URL format for owner param)
+    ref:str=None,  # Git ref (branch/tag/sha) (get from URL not provided); defaults to repo's default branch
+    folder:str=None,  # Only include files under this path (get from URL not provided)
+    show_filters:bool=True,  # Include filter info in title?
+    token:str=None,  # GitHub token (uses GITHUB_TOKEN env var if None)
+    **kwargs  # Passed to `folder2ctx`
+)->Union[str,dict]:  # XML for LM context, or dict of file sizes
+    "Convert GitHub repo to XML context without cloning"
+    import tempfile, tarfile, io
+    if owner.startswith('http'):
+        parsed = parse_gh_url(owner)
+        if not parsed: raise ValueError(f"Invalid GitHub URL: {owner}")
+        owner,repo = parsed['owner'], parsed['repo']
+        ref = ref or parsed.get('ref')
+        folder = folder or parsed.get('path')
+    if repo is None: owner, repo = owner.split('/')
+    api = GhApi(token=token)
+    if ref is None: ref = api.repos.get(owner, repo).default_branch
+    data = api.repos.download_tarball_archive(owner, repo, ref)
+    title = f"GitHub repository contents from {owner}/{repo}/{ref}"
+    if folder: title += f'/{folder}'
+    if show_filters:
+        parts = [f"{k}: {', '.join(v) if isinstance(v, (list,tuple)) else v}" for k,v in kwargs.items() if v]
+        if parts: title += f" (filters applied -- {' | '.join(parts)})"
+    tf = tarfile.open(fileobj=io.BytesIO(data))
+    with tempfile.TemporaryDirectory() as tmp:
+        tf.extractall(tmp, filter='data')
+        subdir = Path(tmp) / tf.getmembers()[0].name.split('/')[0]
+        if folder: subdir = subdir/folder
+        return folder2ctx(subdir, include_base=False, title=title, readme_first=True, **kwargs)

{toolslm-0.3.9.dist-info → toolslm-0.3.17.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: toolslm
-Version: 0.3.9
+Version: 0.3.17
 Summary: Tools to make language models a bit easier to use
 Home-page: https://github.com/AnswerDotAI/toolslm
 Author: Jeremy Howard
@@ -16,8 +16,9 @@ Classifier: License :: OSI Approved :: Apache Software License
 Requires-Python: >=3.9
 Description-Content-Type: text/markdown
 License-File: LICENSE
-Requires-Dist: fastcore>=1.9.6
+Requires-Dist: fastcore>=1.9.7
 Requires-Dist: httpx
+Requires-Dist: ghapi
 Provides-Extra: dev
 Dynamic: author
 Dynamic: author-email

toolslm-0.3.17.dist-info/RECORD

@@ -0,0 +1,13 @@
+toolslm/__init__.py,sha256=HXhmv802-3PQnM5q29vlyrO10zGWKWZQ7xsx3qPYVRM,23
+toolslm/_modidx.py,sha256=EC1pFuHb5MbfRMml7RXx1sxGXlTiczjUimXICuXUMn0,5806
+toolslm/download.py,sha256=yMhyY3u26XRr6a4eZuCCmkprS7LQhHASl01Zn2B4q_o,4481
+toolslm/funccall.py,sha256=_5TyhTjWaWLi-eJ96-4P3_faFv6Ft07nO60UjCF-bPU,11160
+toolslm/md_hier.py,sha256=r_NPezhgfxjRmSYFlu_ND42hXt1qSbaPWHTcjbviOn4,11010
+toolslm/shell.py,sha256=dGInuRKvexu21VmtZkw_0S3BGiTsbAongUG-yG4YHpc,1566
+toolslm/xml.py,sha256=I2lRJPVG6us1g_gOTOKbbnZdlGB3g2-6MYGcLSqkFrM,11173
+toolslm-0.3.17.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+toolslm-0.3.17.dist-info/METADATA,sha256=jSPVU5a6Qe0eYmuxTqj30o0YP81UgHOH_gIAOeeQDaE,2425
+toolslm-0.3.17.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+toolslm-0.3.17.dist-info/entry_points.txt,sha256=xFz0Eymlo5X7BGpaO6DI9gMxvN5A7faebzrlr8ctp5I,95
+toolslm-0.3.17.dist-info/top_level.txt,sha256=4hRTrFWayz_Kz5221XjvlpCwVFrW3WPi1P0fllkTq9s,8
+toolslm-0.3.17.dist-info/RECORD,,

toolslm-0.3.9.dist-info/RECORD

@@ -1,13 +0,0 @@
-toolslm/__init__.py,sha256=xmkmdvq15kb61xdtCoa1YARnvHBnUgI-0GWIJYvHNeA,22
-toolslm/_modidx.py,sha256=kpgsDpj-Tvn90wezrHaMttyzhNcyNVgw_dQgK10qotI,5308
-toolslm/download.py,sha256=g3BxUSxylC_575M7RFSJ1GI3Co3EwPDdEeWzxaf2Czk,4451
-toolslm/funccall.py,sha256=0OBrx6KzI0KK13L-5Hn69yah9oZhgTsKchmMenCoT0A,10421
-toolslm/md_hier.py,sha256=r_NPezhgfxjRmSYFlu_ND42hXt1qSbaPWHTcjbviOn4,11010
-toolslm/shell.py,sha256=dGInuRKvexu21VmtZkw_0S3BGiTsbAongUG-yG4YHpc,1566
-toolslm/xml.py,sha256=tAHoqXrTRiX8i3pR-9KpHoBb8QXJ_TKEVyTEOPviudE,8095
-toolslm-0.3.9.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-toolslm-0.3.9.dist-info/METADATA,sha256=djSwIqYu8Taj8g0yyXKw3IqFr_fbAKhbI3aQu14kv9U,2403
-toolslm-0.3.9.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-toolslm-0.3.9.dist-info/entry_points.txt,sha256=xFz0Eymlo5X7BGpaO6DI9gMxvN5A7faebzrlr8ctp5I,95
-toolslm-0.3.9.dist-info/top_level.txt,sha256=4hRTrFWayz_Kz5221XjvlpCwVFrW3WPi1P0fllkTq9s,8
-toolslm-0.3.9.dist-info/RECORD,,