einlesen_pipeline.svg

def einlesen_pipeline() -> str:
    b = [eyebrow(40, 40, "Von der Rohdatei zum DataFrame")]
    cy = 110
    # Rohdatei
    b.append(rect(40, cy - 40, 190, 118, r=12, fill=T["soft"]))
    b.append(txt(56, cy - 18, "labor.csv", size=12.5, fill=T["ink"], w=600, mono=True))
    for k, ln in enumerate(["patient_id;laktat", "1042;3,4", "1043;1,6"]):
        b.append(txt(56, cy + 6 + k * 20, ln, size=11, fill=T["mut"], mono=True))
    # Parser
    px = 300
    b.append(rect(px, cy - 40, 200, 118, r=12, fill=T["card"], stroke=T["rule"]))
    b.append(txt(px + 100, cy - 18, "einlesen", size=12.5, fill=T["acc2"], w=600, anchor="middle"))
    for k, (lab) in enumerate(["Trennzeichen  ;", "Encoding  UTF-8", "Dezimal  ,", "Header  Zeile 1"]):
        yy = cy + 2 + k * 19
        b.append(txt(px + 16, yy, lab.split("  ")[0], size=10.5, fill=T["mut"]))
        b.append(txt(px + 184, yy, lab.split("  ")[1], size=10.5, fill=T["acc"], w=600, anchor="end", mono=True))
    # DataFrame
    dx = 570
    b.append(rect(dx, cy - 40, 150, 118, r=12, fill=T["card"], stroke=T["rule"]))
    b.append(txt(dx + 75, cy - 18, "DataFrame", size=12.5, fill=T["ink"], w=600, anchor="middle"))
    b.append(rect(dx + 16, cy - 4, 118, 62, r=6, fill=T["soft"]))
    for k in range(3):
        b.append(line(dx + 16, cy + 16 + k * 18, dx + 134, cy + 16 + k * 18, stroke=T["hair"]))
    b.append(line(dx + 62, cy - 4, dx + 62, cy + 58, stroke=T["hair"]))
    b.append(arrow(238, cy + 18, 296, color=T["mut"]))
    b.append(arrow(508, cy + 18, 566, color=T["mut"]))
    b.append(txt(40, cy + 108, "Stolperfallen:", size=11, fill=T["warn"], w=700))
    b.append(txt(140, cy + 108, "„;“ statt „,“  ·  Latin-1 statt UTF-8  ·  Dezimalkomma  ·  fehlender Header",
                 size=11, fill=T["mut"]))
    return canvas(760, 250, "".join(b),
                  "Einlese-Pipeline: Aus einer Rohdatei wird über bewusste Parser-Entscheidungen "
                  "(Trennzeichen, Encoding, Dezimalzeichen, Header) ein sauberer DataFrame.")