cvlac-json 0.1.0__py3-none-any.whl

cvlac_json/__init__.py

@@ -0,0 +1,17 @@
+"""cvlac-json: turn a public CvLAC page (Minciencias, Colombia) into one JSON document.
+
+>>> from cvlac_json import extract
+>>> data = extract("0000026549")            # downloads the page; Scienti ids if an apis.colav.co key is set
+>>> data["profile"]["Nombre"], len(data["works"])
+"""
+
+from importlib.metadata import PackageNotFoundError, version
+
+from .parser import build, extract, fetch, normalize_cod_rh
+
+try:
+    __version__ = version("cvlac-json")
+except PackageNotFoundError:  # running from a source tree without installing
+    __version__ = "0.0.0"
+
+__all__ = ["__version__", "build", "extract", "fetch", "normalize_cod_rh"]

cvlac_json/__main__.py

@@ -0,0 +1,3 @@
+from .cli import main
+
+raise SystemExit(main())

cvlac_json/cli.py

@@ -0,0 +1,79 @@
+"""Command line: ``cvlac-json COD_RH [-o OUT.json] ...``."""
+
+import argparse
+import json
+import sys
+
+from . import __version__
+from .parser import extract, normalize_cod_rh
+
+
+def build_parser():
+    ap = argparse.ArgumentParser(
+        prog="cvlac-json",
+        description="Build a complete JSON from a researcher's public CvLAC page (Minciencias, Colombia).",
+    )
+    ap.add_argument("cod_rh", help="10-digit SCIENTI/CvLAC id (leading zeros are added if missing)")
+    ap.add_argument("-o", "--output", help="output JSON (default cvlac_<COD_RH>.json; '-' for standard output)")
+    ap.add_argument("--html", help="parse this saved page instead of downloading it")
+    ap.add_argument("--save-html", help="also save the downloaded page here")
+    ap.add_argument(
+        "--scienti",
+        choices=["auto", "full", "off"],
+        default="auto",
+        help="auto (default): add the Scienti ids of matching entries when an apis.colav.co key is available; "
+        "full: also add the Scienti records missing from the page; off: public page only",
+    )
+    ap.add_argument("--institution", help="Scienti dump (udea, univalle, uec, unaula; default: first that has the COD_RH)")
+    ap.add_argument("--model-year", help="Scienti model year (default: the latest of that institution)")
+    ap.add_argument(
+        "--apikey",
+        help="apis.colav.co key for this run, overriding $COLAV_APIKEY and ~/.config/colav/apikey "
+        '(it stays in the shell history; prefer --apikey "$(cat FILE)")',
+    )
+    ap.add_argument("--ror", choices=["auto", "off"], default="auto", help="look institutions up in the ROR API (default auto)")
+    ap.add_argument("--version", action="version", version=f"%(prog)s {__version__}")
+    return ap
+
+
+def main(argv=None):
+    a = build_parser().parse_args(argv)
+    try:
+        cod = normalize_cod_rh(a.cod_rh)
+    except ValueError as e:
+        print(f"error: {e}", file=sys.stderr)
+        return 2
+    page = None
+    if a.html:
+        with open(a.html, encoding="latin-1") as fh:
+            page = fh.read()
+    log = lambda msg: print(msg, file=sys.stderr)
+    if page is None and a.save_html:
+        from .parser import fetch
+
+        page = fetch(cod)
+        with open(a.save_html, "w", encoding="latin-1", errors="replace") as fh:
+            fh.write(page)
+    data = extract(
+        cod,
+        html=page,
+        scienti=a.scienti,
+        institution=a.institution,
+        model_year=a.model_year,
+        apikey=a.apikey,
+        ror=a.ror == "auto",
+        log=log,
+    )
+    out = a.output or f"cvlac_{cod}.json"
+    text = json.dumps(data, ensure_ascii=False, indent=1)
+    if out == "-":
+        sys.stdout.write(text + "\n")
+    else:
+        with open(out, "w", encoding="utf-8") as fh:
+            fh.write(text)
+    c = data["counts"]
+    log(
+        f"wrote {out}: {data['profile'].get('Nombre') or '?'} | works {c['works']} (endorsed {c['endorsed_works']}), "
+        f"projects {c['projects']}, events {c['events']}, other sections {len(data['sections'])}"
+    )
+    return 0