Lalabase
Tickets, Doku, Zeiterfassung und Kundengespräche an einem Ort, durchsuchbar per KI. Für Softwarefirmen mit Kundenprojekten. Open Source: auf deinem eigenen Server oder in unserer Cloud in Deutschland.
翻譯準備中
尚未偵測到更新。GitHub 發布版本後會顯示在這裡。
個人 GitHub 帳號下公開儲存庫的擁有者可以登入並驗證所有權,自行管理介紹,無需管理員核准。
登入並產生驗證檔案,在 30 分鐘內將其提交至此儲存庫的預設分支。其他分支或分叉儲存庫中的檔案無效。然後在下方檢查檔案,即可無需管理員核准直接管理專案。請保留檔案以維持管理權限。
如需登入,將前往登入頁面。登入後請回到此處,再次點擊所選的驗證按鈕。
在專案資料夾中執行 /nomorevibe verify,即可驗證所有權並自行管理介紹。
一分鐘完成驗證正在載入完整 README。 以下為已儲存的摘錄。
KISEM Detailed Energy Assessment report generator — IEA, IIT Gandhinagar.
PostMan.html — double-click it. That is everything: the editor, the
report engine, the bill reader, the charts, the exports. Nothing to install,
no server, no npm install, no folder of parts.
To share it with a colleague, send them that one file. To keep a version,
copy it. It opens in any modern browser and keeps your draft in that browser,
so two people opening the same file each have their own work.檢視完整 README ↗KISEM Detailed Energy Assessment report generator — IEA, IIT Gandhinagar.
PostMan.html — double-click it. That is everything: the editor, the
report engine, the bill reader, the charts, the exports. Nothing to install,
no server, no npm install, no folder of parts.
To share it with a colleague, send them that one file. To keep a version, copy it. It opens in any modern browser and keeps your draft in that browser, so two people opening the same file each have their own work.
The same file is served online as index.html (repository
https://github.com/rdudr/PostMAN, deployed by Vercel on every push). Both are
built from the parts in src/ by python src/build.py.
Every keystroke is saved in the browser, as it always was, and the tool runs
with no network at all. Beside that there is now a server at
https://eass.eassiitgn.page — the team's own machine at IIT Gandhinagar,
published through the Cloudflare tunnel already running on it. Sign in and
the report is kept there too, listed for the whole team, with every earlier
save behind it. The server, its API and how to put it on Oracle Cloud are in
server/README.md.
Nothing is lost, and there is no gap between these four rules.
Everybody signs in as themselves, with their KISEM FOX username and password. No second pair to remember, and the activity trail names the person who did each thing rather than a shared account. While a report is open, PostMan watches for a colleague saving it and says so in the strip — "Rahul Patel saved v2", with one button to load their copy — long before anybody can type themselves into a clash.
Opening PostMan asks who you are — every time, on every machine. A reload, a new tab, a laptop picked up the next morning: the sign-in page comes first. Nothing is kept in the browser between sittings, so a shared plant computer holds no session for the next person to walk into. The draft is not touched by this and neither is the report it belongs to: signing in carries on with the same report rather than starting a duplicate, and whatever this browser holds and the server does not goes up the moment there is a password behind it. Within a sitting you stay in — a server that cannot be reached is not a reason to log anybody out, and the session renews itself, so a long day on site does not end in a surprise. Only a definite "that token is no longer valid" returns you to the sign-in page, with the reason on screen and the draft untouched.
Two people on one report do not overwrite each other. The second save is refused, saving pauses rather than retrying into the same wall, and the screen offers the two honest answers side by side: take theirs, or keep mine as a separate report. The editor can tell its own last save from a colleague's, so a reload is never mistaken for a clash.
Nothing, to run. It will reach out for three optional things:
| Report fonts | falls back to close matches if unavailable |
| Spreadsheet reader | needed to import JET-Eff / A-CMP / module workbooks |
| OCR reader | only for photographs of bills; downloads ~5 MB the first time |
Bills supplied as PDFs are read without any of that. If the OCR reader cannot be downloaded — no internet, or a network that blocks it — the screen says so plainly and you read the figures off the bill, which is displayed full-size beside the fields.
Cover & report — company, financial year, logo, gate photograph.
Electricity bills — built to the team's own workbook ("EB Bill & Base", sheet Report Formet): the nine yellow cells per month — month, units, bill amount, power factor, actual MD, energy rate, FPPA rate, night units, TOU units — plus the tariff header (DISCOM, consumer number, contract demand, tariff) are read off the bills; every other column is a formula or a GERC default (85 % minimum billing demand, Rs 150/260 demand slabs, TOU Rs 0.85, EHV 1 %, duty 15 %, the PF rebate/penalty bands). A Difference column reconciles the computed bill against the amount read off the bill, so a misread figure or a default that does not apply shows as red. The full analysis prints on three turned sheets in the workbook's layout — the reader turns the page clockwise — followed by the standard summary, load factors, charts, and the PF and contract-demand proposals. Rates not printed on a bill are derived from its charges and marked so. 2a. Bill capture & verify — drop in bill PDFs or photographs. A DISCOM bill that runs to three sheets stays one bill: the pages of a PDF are grouped by the month they name, the figures are read from whichever sheet carries them, and the others are a click away on the same screen. Figures are read out automatically and shown beside the bill itself; check each one and press Verify. Nothing counts until a person has looked at it. A wrong file is undone with Clear this bill, Delete this month, or Start over.
Import field data — one door for every workbook. JET-Eff, A-CMP, the FOX KISEM export, a PQ-analyser "PostMan export", or the module workbook you can download from that screen and take to site. The file is identified from its own contents and checked against this report's company before anything is written. 3a. Single line diagram — built from cards, not drawn: add a main input, then PCC panels under it, MCC panels and loads under those. Each card holds the rating, kW, kVA, PF, %VTHD, %ITHD and the PQ recording ID; boxes size themselves to their text, a provisional node is dashed, and a node over the IEEE-519 THD limits is outlined red. Click a box to select it and press Delete (or the card's Delete) — its children move up to its parent. 3b. Electrical distribution chapter — assembled in this order: the single line diagram; the demand study; one section per plant main input and then per PCC panel, each with Measured at the panel (the FOX figures) and Power quality analysis (the analyser recording: voltage, current, PF, THD against IEEE-519, harmonic spectrum); the PCC load summary with its charts; MCC panels; the motor load study split into clamp-meter and PQ-analyser readings, each motor marked Working OK, Acceptable, Under-loaded or Overloaded by its load factor; and the APFC panels with a stage-current chart and remarks. Every table fed by FOX carries "Last uploaded data on from ". The full measurement charts for every recording go to at the very end of the report, listed in the Contents.
The ordered list of what is left to do - start the server, point the name at
it, add the accounts, back the folder up - is docs/START-HERE.md.
report design\ the KISEM cover and letterhead artwork as supplied ·
sample bills\ test bills including a 12-month PDF ·
sample reports\ the three reference reports the engine was derived from ·
logos\ IITGN and KISEM marks.
These are source material, not part of the tool. PostMan.html does not read them; you can move them anywhere.
The two feeds: the FOX KISEM app's Excel export (sheets Plant Main
Inputs, PCC Panels, MCC Panels, Motor Loads Clamp, Motor Loads PQ, APFC)
and the AI-PQA analyser. On the analyser's dashboard, Send to
PostMan names the panel and recording ID and parks the recording on
the server for 24 hours; on the Electrical distribution page, List
recordings shows what was sent and Import brings it in. The server
does the heavy work on the whole recording (a week at one second is
fine). Each panel's section then carries the analyser's own charts
(captured from its dashboard, not redrawn), its standards compliance
table (IEEE 519, EN 50160, IEC 61000-3-14, PF — measured, limit, verdict,
remark, score), equipment health (five scored components), the
cost of poor power quality with the tariff figures typed on the
analyser, events detected (counts, the worst events, dips and
swells), data quality and observations; the PCC load summary gains a
Compliance column and Annexure A carries every dashboard chart.
Offline, Save bundle file instead (JSON) or the Excel export drop into
the same box. Recordings are joined to panels by the
Recording ID written on the FOX panel sheet. The contracts are written
up in docs/POSTMAN_EXPORT.md of the AI-PQA repository
(https://github.com/rdudr/AI-PQA) and docs/POSTMAN_IMPORT.md of the FOX
KISEM project.
3c. Boiler and thermic fluid heater from Thermo-X — drop the Thermo-X
app's Excel exchange file (Report → Export Excel) on the Boiler page, the
Thermic fluid heater page or Import field data. The chapter then prints
from it with Thermo-X's own formulas: efficiency summary and chart, the
fuels' laboratory analyses, each boiler's name-plate, the direct method
(daily log, steam table, efficiency, evaporation ratio), the indirect
method (losses L1–L8 per sample and overall, loss chart, the KANE
analyser recording with the samples marked) and the observations. A
Thermo-X boiler typed Thermic fluid heater prints in the TFH chapter.
The typed cards below are used only when there is no Thermo-X data.
Every chapter that takes a workbook has its own drop box, which accepts several files at once; every import merges by the record's own key, so nothing is ever duplicated. How the apps and PostMan are kept in step — and the rule that every touched repository is pushed — is written up in docs/INTEGRATIONS.md.
Modules — switch on only the utilities the plant actually has.
Recommendations — every chapter screen ends with "Add recommendation to this chapter". Each one carries a title, observation and recommendation; optionally a normal and thermal photograph side by side, a graph, and a table of your own figures; then the benefit worked out in the open:
| Technical, electrical | kW saved × hours/day = kWh/day; × working days = kWh/year |
| Technical, thermal | units/hour × hours/day × working days, in kg, litre or scm — one block per fuel, as many as needed |
| Monetary | kWh × ₹/kWh, plus each fuel × its ₹/unit; total is the sum |
| Payback | investment ÷ total annual saving × 12, in months |
The printed page shows every multiplication, so a reader can check it. The certificate, executive summary and savings table all read from the same list, so no two pages can disagree.
Energy, water and GHG baseline — one list of months chosen from menus (never typed) shared by every baseline table. Production in whatever unit the plant counts; Grid Input Energy plus any other electrical sources (solar, turbine); any number of fuels, each with its own unit and calorific value; water from a named source. TOE, specific consumption, totals and averages are all derived from that one set of figures, the conversion factors are printed under every table, and the electrical, thermal, overall and water tables each carry their charts. The GHG chapter's boundary text is standing prose with the figures filled in — Scope 1 per fuel plus mobile combustion, Scope 2 from the grid with on-site generation and renewable offsets shown. Then, as in the sample reports, SCOPE-01 is worked one fuel at a time — a table of month, amount, unit, factor and tCO₂e, then that fuel's chart, then the next fuel — SCOPE-02 the same way for purchased electricity, and an emission baseline with a column per fuel, Scope 2, the total, tCO₂e per unit of production, and its charts. Emission factors are asked for on the GHG screen, one per fuel and one for the grid.
The Average row's specific consumption is total over total for the period, not the mean of the monthly ratios; a low-output month would otherwise pull the year's figure about out of proportion to its energy.
Your own pages — anything the generated sections cannot hold: photos, tables, notes, placed wherever you want in the report. The same screen holds the two print settings: the Contents page (every heading with the page it starts on, each line a link to that page — it stays a link in the PDF the browser saves), and page breaks — every chapter opens a fresh page (that is the default: two chapters on one sheet read as one long section with a bold line in the middle), and on top of that any single topic or sub-topic can open one too, ticked from the report's own outline (chapters, the topics under them, the sub-topics under those, read off the report as it stands). A topic that appears in several places — one per panel, one per machine — breaks at every one of them. Inside one of your own sections, a "Start a new page here" block breaks wherever you like.
Print / PDF — true A4, one page per page.